Skip to main content

Full text of "The Journal of experimental zoology"

See other formats


Journal of Experimental Zoology 






Johns Hupkins University 

Harvard University 


University of Pennsylvania 


Carnegie Institution 

University of Pennsylvania 


University of Chicago 

University of California 

Columbia University 

Harvard University 

University of Chicago 

EDMUND B. WILSON, Columbia University 



Johns Hopkins University 

Managing Editor 







I a. 






2-^ OCT 101 

VVITH THE VOUJf>|IE^^^____ ^ " 

. . . I . n \ i:t.tki;, 
5itive Action of the i;r»ntmii liax^nn llegenerationin Planarians. 


No. 2.— August, 1904. 
Edmund B. Wilson, 

Experimental Studies on Germinal Localization. II. Experiments on the 
Cleavage-Mosaic in Patella and Dentaliimi. With 118 figures 197 

A. J. Carlson, 

Contributions to the Physiology of the Ventral Nerve Cord of Myriapoda 
(Centipedes and Millipedes) . With 6 figures 269 

Frank W. Bancroft, 

Note on the Galvanotropic Reactions of the Medusa Polyorchis penicillata, 
A. Agassiz. With 4 figures 289 

Charles Zeleny, 

Experiments on the I^ocalization of Developmental Factors in the Nemer- 
tine Egg. With 19 figures 293 

T. H. Morgan and Abigail C. Dimon, 

An Examination of the Problems of Physiological "Polarity " and Electrical 
Polarity in the Earthworm 331 

Abigail C. Dimon, 

The Regeneration of a Heteromorphic Tail in Allolobophora foetida 349 





No. 1.— May, 1904. 
Edmund B. Wilson, 

Experimental Studies on Germinal Localization. I The Germ Regions 
in the Egg of Dentalium. With 100 figures 1 

Charles W. Hargitt, 

Regeneration in Rhizostoma Pulmo. With 6 figures 73 

C. M. Child, 

Studies on Regulation. IV. Some Experimental Modifications of Form- 
Regulation in Leptoplana. With 53 figures 95 

T. H. Morgan, 

Self-Fertilization Induced by Artificial Means .• 135 

John Bruce MacC.\llum, 

The Influence of Calcium and Barium on the Secretory Activity of the 
Kidney 179 

Charles Russell Bardeen and F. H. Baetjer, 

The Inhibitive Action of the Rontgen Rays on Regeneration in Planarians . . 191 

No. 2.— August, 1904. 
Edmund B. Wilson, 

Experimental Studies on Germinal Localization. II. Experiments on the 
Cleavage-Mosaic in Patella and Dentalium. With 118 figures 197 

A. J. Carlson, 

Contributions to the Physiology of the Ventral Nerve Cord of Myriapoda 
(Centipedes and Millipedes). With 6 figures 269 

Frank W. Bancroft, 

Note on the Galvanotropic Reactions of the Medusa Polyorchis penicillata, 
A. Agassiz. With 4 figures 289 

Charles Zeleny, 

Experiments on the Localization of Developmental Factors in the Nemer- 
tine Egg. With 19 figures 293 

T. H. Morgan and Abigail C. Dimon, 

An Examination of the Problems of Physiological "Polarity "and Electrical 
Polarity in the Earthworm 331 

Abigail C. Dimon, 

The Regeneration of a Heteromorphic Tail in Allolobophora fcetida 349 

Vernon L. Kellogg, 

Restorative Regeneration, in Nature, of the Starfish Linckia Diplax (Miiller 
and Troschel). With 6 figures 353 

Vernon L. Kellogg and R. G. Bell, 

Notes on Insect Bionomics 357 

No. 3.— November, 1904. 
Florence Peebles, 

The Location of the Chick Embryo upon the Blastoderm. With 2 plates 
and 15 figures in the text 369 

T. H. Morgan, 

Regeneration of Heteromorphic Tails in Posterior Pieces of Planaria 
Simplicissima. With 20 figures 385 

Harry Beal Torrey, 

Biological Studies on Corymorpha. I. C. Palma and Environment. With 
5 figures 395 

Gary N. Calkins, 

Studies on the Life History of Protozoa. IV. Deatli of the A Series. 
Conclusions. With 3 plates and 3 figures in the text 423 

C. M. Child, 

Studies on Regulation. V. The Relation between the Central Nervous Sys- 
tem and Regeneration in Leptoplana: Posterior Regeneration. With 
47 figures 463 

No. 4. — December, 1904. 
C. M. Child, 

Studies on Regulation. VI. The Relation Between the Central Nervous 
System and Regulation in Leptoplana: Anterior and Lateral Regenera- 
tion. With 64 figures 513 

T. H. Morgan and N. M. Stevens, 

Experiments on Polarity in Tubularia. With 5 figures 559 

T. H. Morgan, 

An Attempt to Analyze the Phenomena of Polarity in Tubularia 589 

Vernon L. Kellogg, 

Regeneration in Larval Legs of Silkworms. With 10 figures 593 

Influence of the Primary Reproductive Organs on the Secondary Sexual 
Characters 599 

C. B. Davenport and Marian E. Hubbard, 

Studies in the Evolution of Pecten. IV. Ray Variabihty in Pecten varius . 607 


p. 200, 1. 4: for "vegetable" read "vegetative." 

p. 201, 1. 27: after "both" insert "somatoblasts." 

1.30: omit "CD." 

p. 209, 1. 9: before "pro totrochal" insert "the." 

p. 210, 1. 7: for "circle" read "row." 

p. 212, 1. 4: for "size-relation" read "size-relations." 

p. 217. 1. 7: for "ciliation" read "division." 

p. 236, 1. 14: for "are" read "is." 

p. 237, 1. 13: insert period after "forms." 

1. 34: for " disintegrated" read " disintegrate." 

p. 244, 1. 30: for " essentially" read " in many of its features." 

p. 247, 1. 12: omit "the importance of." 

1.13: omit "of." 

1. 14: for "has" read "have." 

p. 248, 1. 16: for " casual" read " causal." 

p. 253, 1.33: insert hyphen after " entoblast." 

p. 260, 1. 8: for " Ganzbeziehungsweise " read "Ganz- beziehungsweise. 

p. 267, 1. 1: for "Studien" read "Stadien." 





With ioo Figxjres. 

I. Introduction. 

II. Preliminary Observations on the un segmented Egg and the normal 


III. Effect of removing the Polar Lobe. 

(a) General history of the lobeless Embryo, with a Comparison of 

isolated Blastomeres. 

(b) The Mesoblast Question. 

IV. Localization of the apical Organ, and its Correlation with the post-trochal 


V. Localization in the unsegmented Egg. 

(a) Development of Fragments obtained by horizontal or oblique Sec- 


(b) Development of Fragments after vertical Section. 

VI. Observations on enucleated Fragments of fertilized Eggs and on the 

isolated Polar Lobe. 

VII. Comment. 

VIII. Summary. 


The following experimental studies are offered as a contribu- 
tion to the theory of "Organbildende Keimbezirke" or germinal 
prelocalization, especially as applied to the cytoplasmic regions of 
the unsegmented egg. Following the enunciation of the principle 
of "precocious segregation" by Ray Lankester, in 1877, the im- 

iThis work was carried out at the Naples Zoological Station between February 
and August, 1903, on a grant from the Carnegie Institution of Washington, in 
which was included the use of one of the tables subscribed for by the Institution. 
My best thanks are due to the administration of the station for the unfailing 
efficiency and courtesy with which my work was aided in every possible way. 

2 Edmund B. JVilson. 

portance of the cytoplasmic factors of localization and differen- 
tiation was early recognized by Whitman in his remarkable paper 
on Clepsine ('78) and emphasized by him in later papers. Sim- 
ilar views were more or less clearly expressed by Van Beneden, 
Flemming, Platner and others prior to the definite formulation of 
the mosaic-theory of development by Roux in 1888.^ Roux 
himself recognized from the first, as a prominent factor in his 
theory, the importance of a definite topographical grouping of 
specific cytoplasmic materials in the unsegmented egg; though 
unfortunately this was complicated, then and in later discussions, 
by the hypothesis of qualitative nuclear division, which has since 
been shown to be untenable and has now been relinquished by its 
author (Roux, 1903). Since that time the evidence, both cyto- 
logical and experimental, has steadily increased that a prelocaliza- 
tion of the morphogenic factors in the cytoplasmic regions is a 
leading factor in the early development; and it has become evi- 
dent that this is true not only in such "mosaic eggs" as those of 
mollusks or ctenophores, but even in those of echinoderms or 
nemertines, where an isolated blastomere or an egg-fragment may 
produce a perfect dwarf embryo. It has become of high im- 
portance to determine experimentally in what degree such pre- 
localization or cytoplasmic "organization" may exist in the un- 
segmented egg, and to what extent it may vary in different forms. 
It is even more important for our general conception of develop- 
ment to determine by the same method whether the prelocaliza- 
tion of the morphogenic factors, in whatever degree it may occur, 
exists from the beginning, or whether, as the cytological evidence 
seems to show, it is established by a progressive process; for in 
the latter case, as is hardly necessary to point out, prelocalization, 
even in the unsegmented egg, may be brought under the category 
of epigenetic phenomena ("epigenetic qualities" as distinguished 
from "preformed qualities"^), and falls into harmony with hy- 
potheses that assume the nucleus to be the primary determining 

The present studies, which are a continuation of the preceding 

1 Cf. my work on The Cell. 

2 Boveri ('03 ), p. 356. 

Experimental Studies on Germinal Localization. 3 

ones on the nemertlne egg (Wilson, '03) bear upon both these 
questions. In that paper I approached especially the second ques- 
tion in an experimental study of the egg of Cerehratulus, which 
has since been extended by the work of Yatsu ('04). My 
results clearly showed that in this egg the cleavage-factors are 
not definitely localized until after the completion of the ma- 
turation of the egg, but they gave no definite evidence regarding 
the localization of the morphogenic factors (as distinguished 
from those of cleavage) at this period; it was, however, shown 
that in the comparatively young blastula, before the formation of 
the mesoblast, morphogenic localization, as shown in the pre- 
determination of the gut and apical organ, has become much 
more definite than in the unsegmented egg. Yatsu subsequently 
obtained evidence, in the same species, that the localization of the 
morphogenic factors is a progressive process even in the stages 
preceding cleavage, since the percentage of normal larvae ob- 
tained from egg-fragments at successive periods steadily dimin- 
ishes from the first discharge of the eggs (when maturation be- 
gins) up to the period immediately preceding the first cleavage; 
and the nature of the defective larvae, correlated with the plane 
of section, pointed to a increasingly definite localization, in the 
later stages preceding cleavage, of the bases of several important 
organs, such as the apical organ, gut, and ciliated lobes of the 
pilidium. I am now able to offer an experimental analysis along 
the same lines — perhaps I should say the beginning of such an 
analysis — of the molluscan egg, in which pure observation of 
the cell-lineage has produced such convincing evidence of mosaic 
development, sustained by Crampton's initial experimental exam- 
ination of the gasteropod egg ('96), and by the interesting cyto- 
logical work of Lillie ('01) and Conklin ('02) on the cyto- 
plasmic regions of the unsegmented and segmenting egg. The 
cytological and experimental results coincide in demonstrating 
in this egg (specifically in Dentalium) the existence of a very 
definite prelocalization of some of the most important factors 
both of cleavage and morphogenesis, which here closely coincide. 
They show conclusively also, contrary to what the nemertine 
experiments had led me to expect, that in its main features this 

4 Edmund B. JVilson. 

prelocalization exists In the egg at the time It leaves the ovary, 
and probably much earlier, and long before even the Initial 
stages of maturation and fertilization. Nevertheless, progressive 
changes take place during and subsequent to maturation, which, 
when compared with those occurring In other forms, show this 
egg, as I believe, to be only the extreme of a series that connects 
it with such forms as the nemertlne or echlnoderm, and brings 
them under one point of view. 

The present paper deals mainly with the development of frag- 
ments of the unfertilized egg of DentaUum, the eggs being cut 
singly with the scalpel under the microscope and subsequently 
fertilized, following the method of Delage ('99). I shall here 
consider the development of isolated blastomeres only incidentally 
for the sake of comparison, reserving a fuller account for a second 
paper. It may be stated here, however, that the experiments on 
this part of the subject demonstrate, even more conclusively than 
do those of FIschel for the ctenophore-egg, that the cleavage of 
the ovum, In both DentaUum and Patella, Is In fact what the 
normal cell-lineage so clearly indicates, essentially a mosaic-work, 
In accordance with Crampton's earlier experiments on Ilyanassa. 
Blastomeres Isolated at any stage from the 2-cell onward con- 
tinue to segment as if still forming part of a complete embryo; 
and apart from the changes due to shifting of the cells, which, as 
in the ctenophore, often lead to the displacements of the larval 
structures and to the closing of the partial embryos, undergo 
essentially the same differentiation as If united to their fellows. 
Thus, the first two blastomeres, upon separation, give rise to two 
dissimilar larvae, each of which Is defective and represents es- 
sentially the same structures as would have been produced had the 
two cells remained united; In hke manner, of the isolated cells 
of the 4-cell stage, the larva from the D-quadrant possesses cer- 
tain structures that are lacking in the other three; and the dif- 
ferences among the larvae from cells of the 8- or i6-cell stages are 
still greater. Cells procured by successive Isolations up to the 
64-cell stage, or later, differentiate singly, according to their na- 
ture, into actively swimming trochoblasts of three kinds; into 
ordinary ectoblast- or entoblast-cells, into sensory cells bearing 

Experimental Studies on Germinal Localization. 5 

the characteristic sensory hairs of the apical organ; and even into 
what I believ^e to be muscle-cells and mesenchyme-cells, though, 
unlike the foregoing cases, the precise origin of these was not 
traced. These eggs thus represent the opposite extreme to such 
forms as those of Amphioxiis, the echinoderm, or the nemertine, 
and give a result which, apart from the hypothesis of qualitative 
nuclear division, agrees essentially with Roux's original conception 
of mosaic-development, with the conclusions of many students 
of cell-lineage, with the experimental results of Crampton on 
the gasteropod-egg, and with those of Fischel regarding the 



The egg of Dentalium, like that of Cerebratuliis, possesses 
certain features by means of which the axis may be determined 
in the living egg from the moment of its release from the ovary. 
The egg is more or less deeply pigmented, perfectly opaque, and 
of a color that varies in different individuals from light olivaceous 
to reddish brown or almost brick red. When first set free the 
egg is somewhat irregular, but quickly becomes more rounded. 
It is then seen to be very considerably flattened, so as often to 
be almost biscuit-shaped, one side being always more flattened 
than the other, and often more or less Irregular In contour. 
Viewed by reflected light the central region of each of the flat- 
tened sides is seen to be occupied by a very distinct, though vaguely 
bounded, white area, nearly or quite free from pigment (Fig. i) ; 
these areas, as shown by the subsequent development, correspond 
with the two poles of the egg, and the more flattened side, which 

1 The eggs of Patella, which were employed mainly for a study of the iso- 
lated blastomeres, were available from the middle of March until the latter 
part of May. Those of Dentalium, which were used especially for the develop- 
ment of egg-fragments, first became mature at the beginning of June, when 
less than two months remained for their study. The shortness of this period 
accounts for some of the obvious gaps in my work. The complexity of the 
subject, and the practical difficulties presented by the material are such that 
more extended work, with additional material, will be required for its com- 

Edmund B. JVilson. 

is the side of attachment in the ovary, is found to represent the 
lower or vegetative hemisphere. 

Fig. I. 
Cleavage, from living Eggs. 
I, Outline of egg soon after release, in polar view, showing white polar area; 
2, the same egg, 20 minutes later, after throwing off the membrane; 3, similar 
egg, from the side ; 4, egg one hour after fertilization, with fertilization-mem- 
branei and polar bodies; 5, beginning of the first cleavage, formation of the 
polar lobe; 6, trefoil, i^ hours after fertilization; 7, resulting 2-cell stage; 
8, beginning of second cleavage from the side, second polar lobe forming; 9, 
second cleavage at its height. 

1 Accidentally omitted by engraver. 

Experimental Studies on Germinal Localization. 7 

During the 20-30 minutes following Its release the ripe, un- 
fertilized egg becomes nearly spherical (and hence appears con- 
siderably smaller In polar view), the membrane by which it is 
at first surrounded separates more widely from the egg, finally 
ruptures suddenly, and then quickly draws together at one side, 
where It Is thrown off as a mass of debris attached to the egg 
(Fig. 2).^ Following this, a substance which at first surrounds 
the egg as a thin, transparent layer swells up to form a jelly, which 
raises the egg slightly from the bottom. The wall of the ger- 
minal vesicle breaks down at about this period (20-30 m.), leav- 
ing a clearer space In which the first maturation-figure appears. 
The white polar areas are still clearly visible, and the egg, still 
unfertilized, now gives the appearance of being surrounded by a 
very broad, horizontal pigment-ring, which, though often faint 
and with vague boundary, is always distinctly visible (Figs. 3, 
4). The ring recalls that described by Boverl ('01) in the egg 
of Strongylocentrotiis , though relatively broader. The egg of 
Dentaliiim thus shows a visible stratification of material analo- 
gous to the zones seen In Strongylocentrotiis ; but, unlike the lat- 
ter, the zones of Dentaliiim clearly pre-exist before even the pre- 
paratory changes of maturation take place. 

Sections and total preparations of the flattened egg, fixed shortly 
after Its discharge or removal from the ovary, show that a distinct 
structural modification exists in each of the white areas, at this 
period much more marked in case of the lower or vegetative area. 
Surrounding the lower pole (Fig. 10) is a very distinct mass of 
dense almost homogeneous protoplasm, of approximately the same 

1 All the figures were outlined as accurately as possible with the camera, and 
with the exception of Figs. 10-13 and 33, 38-41, are enlarged to the same scale 
(150 diameters). They are only schematized in that the pigment is represented 
by stippling, whereas the color does not actually appear in the form of dis- 
tinct granules, but as a nearly uniform hue. The stippling somewhat exag- 
gerates the distinctness of the pigment as seen in most individuals; though in 
the most deeply pigmented ones, viewed under strong direct light, the color 
appears with great distinctness and its limits may be clearly seen. The opera- 
tion of cutting usually leads to disturbances in the arrangement of the pig- 
ment, so that frequently no definite color-pattern can be clearly made out in 
the dwarf embryos. I have only represented the pigment in cases where its 
boundaries could actually be seen. 

8 Edmund B. Jf'ilson. 

extent as the white area seen in the living egg; this contains no 
yolk-spheres, and stains with great intensity with a strong plasma- 
stain like Congo red. This mass, sharply marked off from 
the surrounding yolk, bulges slightly outward at the surface and 
at the margin is continuous with a very thin ectoplasmic zone that 
entirely surrounds the egg, but is only clearly visible in sections. 
Internally this mass is confluent with a somewhat narrow zone of 
similar finely granular protoplasm that extends upwards partly 
around the germinal vesicle. It is probably to the presence of 
this remarkable protoplasmic mass that the appearance of the 
lower white area is due, though the latter may have a different 
cause. In a general way, the lower protoplasmic area is un- 
doubtedly comparable with the lower zone, composed of green 
material, seen in the egg of Myzostoma (Beard, Wheeler, and 
Driesch), as is proved by its later history. Comparison of my 
Fig. lo with Wheeler's Fig. 2 ('97), will show how closely 
similar the relations of the lower protoplasmic area in the two 
eggs are.^ 

The upper white area cannot be distinguished as such in the 
fixed eggs, and is apparently produced by a different cause from 
the lower one. Exactly at the upper pole is a very small, super- 
ficial disc of clear, dense, intensely staining protoplasm, which, 
like the lower protoplasmic mass, is continuous at its margin 
with the general ectoplasmic layer (Fig. 10), This upper 
disc is so small as readily to escape observation; but suf- 
ficiently careful examination invariably reveals its presence, which 
is furthermore frequently indicated by a slight indentation of 
the egg-periphery at this point. It varies considerably in thick- 
ness and extent in different specimens, but is always very small 
at the beginning.^ Evidently, the upper protoplasmic disc is not 
large enough to account for the appearance of the upper white 
area in the living egg, which must be due to some other cause. 

1 Compare also Driesch, '96, Fig. 12. 

2 Sfctions of the ovary show that both the upper disc and the lower proto- 
plasms area are present while the egg is still attached to the ovarian wall. The 
eggs are greatly distorted in shape, but in a general way are pyriform, and at- 
tached by the narrow end. The lower protoplasmic area occupies the narrower 
end, by which the eggs are attached ; the upper disc is at the opposite point. 

Experimental Studies on Germinal Localization. 9 

perhaps to a lighter tint In the deutoplasm In this region. In the 
following account, accordingly, it will be necessary always to 
distinguish clearly between the upper white area, or polar area, 
and the upper protoplasmic disc or area. 




12 13 

Fig. II. 

Vertical Sections of the Normal Egg. 

Fig. 13 directly from section (picro-acetic) ; outlines of Figs. 10-12 (sublimate- 
acetic) from optical section of total preparations, details from actual sections. 
The peripheral zone of deeply staining yolk shown in Fig. 13 occurs in all these 
stages after picro-acetic fixation, but not after sublimate-acetic. 

10, Unfertilized egg, five minutes after release, showing both protoplasmic 
areas; chromosome-like bodies in the nucleolus; 11, fertilized egg, 30 minutes alter 
-fertilization , first polar spindle ; 12, fertilized egg, 60 minutes after fertilization , 
initial stage in formation of polar lobe; 13, first cleavage, 68 minutes after fer- 
tilization, just before the complete trefoil stage. 

10 Edmund B. Wilson. 

I shall here give only a very general account of the later his- 
tory of the two protoplasmic areas, which will require a thorough 
cytological study for its full elucidation. As the egg, still unfer- 
tihzed, lies in sea-water, the ectoplasm in the region of the upper 
disc slowly increases in amount, and in some cases this region 
shows a faintly radiating appearance around its periphery as If 
clear hyaloplasm were flowing into it from the surrounding region. 
I am uncertain whether in this process the original disc itself 
enlarges or is only surrounded by an accumulation of hyaloplasm 
— a point of Importance for the comparison with the upper polar 
ring of the annelid egg that is drawn further on. I shall continue 
to speak of the ectoplasmic thickening at the top of the egg as the 
"upper protoplasmic area," but would call attention especially 
to the fact that the original disc is composed of very dense homo- 
geneous protoplasm that differs markedly in character from th*: 
alveolar protoplasm of the ectoplasmic thickening that afterwards 
extends over the whole upper surface of the egg.^ 

When the germinal vesicle breaks down, the maturatlon-splndle, 
which is relatively small. Is formed just below this protoplasmic 
area, rotating into a radial position and moving towards the 
periphery so that Its outer end lies in or just below It (Fig. ii). 
In this position it remains, in metaphase, until the egg is fertilized, 
when the divisions proceed, the polar bodies being successively 
extruded exactly at the upper pole, at the centre of the upper pro- 
toplasmic area (which Is now rapidly extending and shows no defi- 
nite boundary), and hence at the centre of the upper white area 
(Fig. 4). At this period the protoplasmic area comes Into con- 
nection by a rather narrow neck of hyaloplasm, in which the spin- 
dle lies, with the central mass left after the germinal vesicle breaks 
down. After the polar bodies are formed this connection is sev- 
ered, and the upper protoplasimc area spreads out still more 

1 The general ectoplasmic layer can in the earlier stages hardly be seen in 
total preparations, but appears clearly in sections either after staining with 
haematoxylin and a strong plasma-stain such as Congo red (when it appears 
clear red) or after borax carmine. It is at first much thinner and less defi- 
nitely bounded than, for instance, in Rhynchelmis as figured by Vejdovsky,'88 (in 
the recent paper of Vejdovsky and Mrazek, '03, it is represented as much thinner 
than in the earlier paper), but later becomes very conspicuous. 

Experimental Studies on Germinal Localization. 1 1 

widely so as to appear as a general thickening of the ectoplasmic 
layer over the whole upper hemisphere (Figs. 12, 13). This 
thiclcening is most marked near the animal pole, where it is very 
conspicuous at the time of cleavage, extending thence approxi- 
mately to the equator of the egg, or slightly below it, but without 
any very definite margin. It stains deep red in Congo red and 
shows a finely alveolar structure quite unlike that of the original 

During the foregoing stages marked changes occur also in the 
lower protoplasmic area, and it is evident that active movements 
of its material take place. These are perhaps due in part to the 
entrance of the spermatozoon at the lower pole, but in part also 
to the fact that upon the breaking down of the germinal vesicle 
the finely granular material derived from It becomes more or less 
definitely confluent with the lower area (as Wheeler describes 
in Myzostoma) , so that an irregular pillar of protoplasm, sur- 
rounded on all sides by yolk, now extends from the lower pole 
nearly to the upper protoplasmic area (Fig.ii) and ultimately 
becomes connected with the latter as the first maturation spindle 
moves upwards.^ In vertical section it may very clearly be seen 
that the material of the upper part of this pillar differs markedly 
from the lower, both In texture and in staining capacity (the two 
regions show a rather distinct boundary, indicated by the dotted 
line in Fig. 11), the lower region being very dense and staining 
in the double stain clear red, the upper one much looser (alveo- 
lar?) in structure and staining purple or blue. During the polar 
body formation the lower area changes its form, often becoming 
irregular and sometimes elongate or sickle-shaped. It is a note- 
worthy fact that at the time each polar body Is extruded the egg 
becomes irregular in contour or almost amoeboid, at the center 
of the loiier polar area, afterwards resuming Its even outline.^ 
After formation of the polar bodies the upper part of the proto- 
plasmic pillar retreats from the periphery, while the yolk again 
extends across the upper region above the egg-nucleus. In the 
upper part of the internal protoplasmic region conjugation of the 

1 Cf. Wheeler's Fig. 10 or 16. 

2 This was figured by Lacaze Duthiers ('57 ) nearly fifty years ago. 

1 2 Edmund B. JVilson. 

germ-nuclei takes place. At the period shortly preceding the 
first cleavage, when the upper disc has been replaced by the very 
broad ectoplasmic thickening described above, the lower proto- 
plasmic area, as seen in surface views of total preparations, varies 
a good deal in appearance in different individuals, being some- 
times rounded and fairly well circumscribed, sometimes irregular, 
or even broken up so as to present a mottled appearance. 

The first cleavage, which occurs about thirty minutes after the 
extrusion of the second polar body, is characterized by a trefoil 
stage, like that occuring in many gasteropods, lamellibranchs 
and annelids (Figs. 5, 6). Exactly surrounding the lower pole 
is formed, by a horizontal constriction, a large lobe, into which 
passes the whole of the lower white polar area, and M^hich, like 
the area itself, appears pure white in the living object. Since 
the surface of the lobe is much larger than that of the original 
lower polar area from which it arises, it is evident that material 
from the interior of the egg must How into the lobe as it form.s. 
Vertical sections of the egg as the polar lobe begins to form show 
somewhat varying appearances, due in part to differences in the 
plane of section, but also in part to varying conditions in the 
protoplasmic area itself. The rather small cleavage-figure, at 
this period entirely surrounded by deutoplasm, lies in late ana- 
phase or early telophase slightly above the centre of the egg. 
At the lower pole the dense protoplasm of the lower area is now 
spread out, more or less irregularly, to form a thick peripheral 
layer that fades away insensibly into the yolk-bearing region. 
Frequently, as in Fig. 12 {cf. Wheeler's Fig. 46) this thickening 
appears fairly regular and symmetrical and suggests the ecto- 
plasmic thickening that precedes the formation of a pseudopod 
in Amceha; sometimes it is less regular than this, and occasion- 
ally gives the appearance of an asymmetrical wedge-shaped mass 
extending into the yolk. As the lobe forms it receives this clear 
protoplasm, accompanied by an inflow of yolk that seems to in- 
vade the clear substance more or less; so that in section scattered 
yolk-granules are found in the lobe and frequently no definite 
boundary of the clear substance can be distinguished (Fig. 13). 
In any case it is certain that the whole of the lower protoplasmic 

Experimental Studies en Germinal Localization, 13 

area passes into the lobe (like the green material of the Myzos- 
tovia egg) to constitute its main bulk, precisely as Wheeler shows 
in Myzostoma {cf. his Fig 47). The term "yolk-lobe" em- 
ployed by a number of earlier observers is therefore as mislead- 
ing as it is inappropriate and may be replaced by the term "polar 
lobe." For reasons given in the discussion at the end, I believe 
it very probable that at least the lower protoplasmic area, and 
probably also the upper disc, are in a general way comparable to, 
if not identical with, the polar rings observed in the eggs of cer- 
tain leeches and oligochaetes. 

Immediately after the polar lobe is formed a vertical furrow 
cuts into the egg from the upper pole, dividing the upper white 
area into equal parts and forming with the polar lobe a trefoil, 
of which the two upper lobes are of exactly equal size and contain 
all of the pigment, while the unpigmented polar lobe is consider- 
ably less than half the bulk of each of the others (measurements 
give a ratio of i to 0.32-0.46, Fig. 6) . At the height of its form- 
ation the trefoil appears at first sight to consist of three separate 
spheres. Close examination invariably shows however that the po- 
lar lobe Is united to one of the upper lobes by a very narrow pedicle 
which is never severed; and as the cleavage proceeds these two 
lobes completely fuse while the remaining upper lobe Is cut off as 
a separate blastomere. Thus Is formed a characteristic unequal 
2-cell stage (Fig. 7), consisting of a smaller anterior cell, AB, 
and a larger posterior one, CD, which differ In volume by ex- 
actly the bulk of the polar lobe. Each of these cells has at the 
upper pole a white area, representing half the original upper 
polar area. The lower polar area, on the other hand, is con- 
fined to the larger cell, and obviously represents that part of the 
substance of the fused polar lobe that appears at the surface, a 
part having again moved into the interior of the egg.^ Upon 
the 2-cell stage thus formed is moulded the entire subsequent 
development, which in Its general outline Is of essentially the same 
type as In such forms as Unio or Nereis. 

The experiments recorded In this paper relate mainly to the 
significance of the material of the lower polar area, and of the 
polar lobe, and form a continuation of those begun by Crampton 

1 Cf. Wheeler's Fig. 


Edmund B. Wilson. 

in his interesting experimental paper on Ilyanassa, published in 
1896. In order to understand the significance of the experiments 
to be described it will be necessary to trace briefly the subsequent 
development. The second cleavage is ushered in by the reap- 
pearance of the polar lobe at the vegetative pole of the larger 
cell, CD, of the same size and form as before, and again consist- 
ing entirely of white material (Figs. 8, 9). The cleavage in 
this cell, whether separated from its fellow or remaining united 

Fig. III. 
Cleavage, from living Eggs. 
14, Four-cell stage, from lower pole; 15, beginning of third cleavage, from 
lower pole, third polar lobe; 16, eight-cell stage, from lower pole; 17, beginning 
of fourth cleavage , first somatoblast in formation ; 18, sixteen-cell stage, from 
lower pole; 19, view from lower pole, after the formation of the third quartet; 
19a, D (pigmented) and 4d, immediately after division ; surface view. 

with it, follows the same general course as in the first cleavage of 
the entire egg, the polar lobe finally fusing with one of the cells, 
namely, D, the left posterior quadrant, where it again forms a 
very definite lower polar white area.^ The anterior cell, AB, 

1 Cf. Wheeler's Fig. 49, Driesch's ('96 ) Fig. 12, of Mysostoma. 

Experimental Studies on Germinal Localization. 15 

in the meantime divides equally, without the formation of a polar 
lobe. In the 4-eell stage, accordingly, the large posterior cell, 
D, exceeds A, B or C, by exactly the volume of the lobe, and 
the lower white area appears only in D (Fig. 14). On the other 
hand, the substance of the original upper white area is equally 
distributed among the four; but it is evident that the amount of 
white material visible at the surface has somewhat increased. 
The 4-cell stage shows the characteristic relations of the blasto- 
meres observed in so many other eggs of this type. The two 
lateral cells, A and C, He at a higher plane, and are In contact 
along the upper side by an upper "cross-furrow." B and D, on 
the other hand, are in contact along a longer transverse lower 
cross-furrow; and these characters, together with the large size 
of the posterior cell, D, thus give an immediate means of orienta- 
tion from this time forwards. 

As the egg prepares for the third cleavage the upper white 
material shifts slightly towards the left upper angle in each quad- 
rant, anticipating the formation of the first quartet of ectomeres 
by the usual dexlotropic cleavage. These cells, which are of 
equal size and In the A, B and C quadrants are not much smaller 
than the basals, are formed entirely from the white material of 
the upper polar areas; and it Is here again evident that an ex- 
tensive flow of this material must take place from the interior 
of the egg. Their formation does not, however, exhaust the 
white substance of the upper areas, which still remain In the 
upper regions of the four basals. During this division the polar 
lobe forms for the third and last time, from the white material 
of the lower area, in the D-quadrant; but it Is now noticeably 
smaller than before, and does not constrict so deeply (Fig. 15). 
After the completion of the cleavage the lobe again fuses with 
D, in which, as the egg enters Into the "resting stage," the lower 
white area still appears; though this soon undergoes a great 
change (Fig. 16). 

The fourth cleavage is of especial interest, since a large part 
of the substance of the lower white area now passes into the first 
somatoblast, id, or X, and Is thus for the first time actually cut 
off from the pigmented region. This cleavage is preceded and 

1 6 Edmund B. Wilson. 

accompanied by an extensive shifting of the cytoplasmic mater- 
ials in all of the cells. In the three basals, A, B and C, the white 
material towards the animal pole moves over towards the upper 
right angle of the cell and Increases in amount, extending so far 
down the egg that in some individuals it may be seen, when the 
egg is viewed from the vegetative pole, as a narrow white cres- 
centic area (Fig. i6). A similar process takes place in D, but 
in addition to this a great change takes place in the white material 
of the lower polar area, which leaves its position at the lower pole, 
moves over towards the same side as the upper white area, and 
finally fuses with it, while the pigmented part becomes lighter in 
color, often irregular or mottled in appearance, and extends into 
the area formerly occupied by the lower white substance. In 
the ensuing cleavage, D is usually the first to divide, giving 
rise by a leiotropic cleavage to the large first somatoblast, 2d or 
X (Figs. 17, 18). This cell consists almost entirely of white 
material which is certainly derived in large part from the orig- 
inal lower white area, but undoubtedly also in part from the upper 
white area, which, as stated above, fuses with the lower area in 
the period preceding this cleavage. In some cases X receives 
also a small amount of the pigment (Fig. 18), in others it seems 
to be composed entirely of white material. The other members 
of the second quartet, 2a, 2b, and 2c, are much smaller than X, 
and each is formed mainly from the white material of the upper 
polar area, but as a rule, perhaps always, each receives also a 
variable amount of pigment. During the foregoing changes the 
upper quartet divide leiotropically in the usual fashion, to form 
the four primary trochoblasts, which are slightly smaller than 
the upper cells. Owing to the foregoing changes the pigment, 
which in the unsegmented egg extended far up towards the animal 
pole, has been moved downwards so as to lie below the 
equator of the egg, most of it being contained in A, B and 
C, some in D, a little in 2a, 2b and 2C, and sometimes also a little 
in 2d. The pigment becomes still more restricted during the 
fifth cleavage, since the micromeres of the third quartet are again 
mainly composed of white substance. 

Experimental Studies on Germinal Localization. 17 

The fifth cleavage, dexlotropic in all the cells, produces the third 
quartet, each cell of which is considerably smaller than the corre- 
sponding basal (Fig. 19). Qf these cells 3d is much the largest, 
and is usually composed entirely of white material, while 3a, 3b 
and 3c usually, perhaps always, receive a certain amount of 
pigment. At the end of the cleavage the macromeres rapidly 
diminish in apparent size, evidently owing to their passing more 
deeply into the egg, and the color-pattern becomes more or less 
confused, though A, B and C still show the greatest amount of 
pigment, while D distinctly shows a white area on the side turned 
towards X, where 4d is subsequently formed. I have not been 
able to observe the formation of the entire fourth quartet satis- 
factorily, either in the opaque living object or in preparations. 
I can however state positively that as seen in surface-view of the 
living egg, 4d is very small (smaller than 3d and very much 
smaller than 2d) and appears pure white (Fig 19, a). I have 
been unable to determine whether the white material of this cell 
is derived from that of the original lower white area; though, 
as will appear hereafter, the experimental evidence indicates that 
such is the case. At this period the four basals appear much 
smaller, having evidently retreated into the interior. 

Beyond this point it is not necessary at this time to trace the 
cleavage. The foregoing observations clearly show that, in Den- 
talium the freshly discharged egg, prior to maturation or fer- 
tilization, shows a definite segregation of zisibly different ma- 
terials which accurately foreshadows a corresponding distribution 
of these materials among the hlastomcres during cleavage. Of 
the three zones of material superficially visible in the living egg, 
the upper one (upper white area) is allotted to the first three 
quartets of ectomeres, apparently in equal amount in each quad- 
rant; the middle pigmented zone is mainly allotted to the four 
basal entomeres, though a portion also passes into ectomeres of 
the second and third quartets; while the lower zone (lower white 
area) certainly passes mainly into the first somatoblast, 2d, or 
X, probably in part into the second somatoblast, 4d, or M, 
and possibly in part into the left posterior micromere, 3d, of the 
third quartet. This agrees in general with the history of the 

1 8 Edmund B. Wilson. 

zones visible in the living egg of Myzostoma, as observed by 
Driesch ('96), where the lower polar area is represented by a 
green substance, the upper one by a reddish material, and the 
pigment zone of Dentalium by a zone of clear protoplasm. It 
is important not to confuse the above-described distribution of 
white and pigmented material with that of protoplasm and deu- 
toplasm. As shown on a preceding page the upper white area is 
not, like the lower one, free from yolk; and in point of fact all 
the cells contain a large amount of yolk. The pigment-pattern is 
only a visible expression in the living object of a distribution of 
specific materials that can only in part be distinguished in sections. 
We may now briefly consider the main outlines of the larval 
development. In warm weather the embryos become ciliated at 
about the ninth or tenth hour, and at the end of twenty-four hours 
are well developed trochophores that swim very actively at the 
surface, progressing in a spiral curve and rotating from right to 
left as seen from the side. At this period (Fig. 29) the body is 
of a blunt spindle-shape, encircled at the equator by a very broad 
prototroch composed of three principal rows of large trocho- 
blasts which bear three corresponding rows of powerful cilia 
completely encircling the body and leaving no dorsal gap (as is 
also the case in Patella). The pre-trochal and post-trochal re- 
gions, while somewhat variable, are at this period nearly similar 
in form and size, being roughly conical and rounded at the tip. 
The pre-trochal region is wholly covered with very short vibra- 
tile cilia and bears at its apex a very long and well-defined tuft 
of flexible, but not vibratile, flagelliform sensory hairs. In total 
preparations, or in longitudinal sections, it may be seen with 
great clearness that the apical tuft is borne upon a large and 
definitely circumscribed apical thickening or plate, sharply marked 
off from the surrounding cells. The post-trochal region is not 
ciliated, but bears at its posterior extremity a small bunch of 
sensory hairs, which differ from those of the apical tuft in being 
quite stiff, and radiating from the common point of attachment. 
The ahmentary canal at this period forms a closed sac divided 
into two chambers, into one of which at a slightly later period 
opens the mouth, formed immediately below the prototroch, but 

Experimental Studies on Germinal Localization. 19 

the anus does not yet exist. The post-trochal region already shows 
the mantle fold and the beginning of the shell-gland. On either 
side the gut may be seen an irregular mass of small cells which 
I believe to represent the coelomesoblast, though I have not yet 
traced them to the pole-cells. These masses are not to be con- 
founded with two masses lying further forward that are pro- 
liferated off from the ectoblast in two symmetrically placed lateral 
areas in the pre-trochal region and perhaps represent a part of 
the paedomesoblast (ectomesoblast) or perhaps the foundations 
of* the cerebral ganglia. These areas, which are figured by Ko- 
welevsky ('83, Figs. 32, 37, 55) are shown in the lobeless em- 
bryos (Figs. 33, 40). 

The ensuing changes take place very much more rapidly In 
the Naples species (D. entalis) than in the northern form stud- 
ied by Lacaze Duthiers ('57), which is probably due in a meas- 
ure to the higher temperature. By the 30th hour the post-trochal 
region has considerably elongated and the pre-trochal region is 
somewhat diminished (Fig. 30). In the course of the ensuing 
twelve hours the pre-trochal region wholly disappears from view, 
being withdrawn into the interior, while the post-trochal region 
becomes still more elongated and the larva sinks to the bottom, 
where it swims only sluggishly. About this time the body becomes 
surrounded by .an extremely delicate hyaline shell into which 
the greatly diminished prototroch can be withdrawn; and by 
the end of the second day the foot appears on the median ventral 
side. By the end of the third day the foot has become a large 
protrusible organ, trilobed towards the free end, and the pro- 
trotroch is still smaller (Fig. 31, which closely agrees with 
Lacaze's Fig. i, Plate VIII). In many cases the metamorphosis 
is complete by the end of the fifth day, the prototroch having 
disappeared, the otocysts and pedal ganglia being clearly visible, 
and the young Dentaliiim assumes the condition figured by Lacaze 
on Plate 8, Figs. 2, 3 — a larva of 20-25 days ( !). 

Many details have been omitted from the above account that 
have already been described In the well-known memoirs of Lacaze 
Duthiers ('57) and Kowalevsky ('83). Many others will re- 
quire for their full elucidation much more extended study than 

20 Edmund B. Wilson. 

I have thus far been able to devote to the subject. The greatest 
gap in my work thus far is the failure to trace the connected 
history of the mesoblast, which can only be done by a complete 
study of the cell-lineage. This presents considerable obstacles 
owing to the difficulty of obtaining good total preparations at 
every stage (the eggs and embryos stain diffusely in most dyes, and 
the great abundance of deeply staining yolk in all the cells renders 
it difficult to get clear pictures) , and my time was so taken up with 
the study of the living material that I had not opportunity to 
work out a really satisfactory method. For sectioning the best 
results were given by sublimate-acetic, the sections being stained 
with thionin, which gives a sharp nuclear stain without coloring 
the yolk. The best total preparations were obtained by mount- 
ing in balsam without staining. Apart from the technical diffi- 
culties, the object is Itself difficult, in the earlier larval stages on 
account of the difficulty of distinguishing between mesoblastic 
and entoblastic elements in the crowded mesentoblast-mass, in the 
later ones by reason of the complication introduced by the folding 
of the mantle and the shell-gland. 


The ease with which the eggs of Dentaluim may be operated 
recalls the remark of Lacaze that "L'embryon du Dentale est un 
de ces exemples f aits pour I'etude du developpement" ('57, p. 196). 
For experimental purposes however it presents certain difficulties 
that should carefully be borne in mind in considering the results 
of the operations. First, there is a certain amount of variation, 
not wide but still noticeable, in the size of the eggs and the re- 
sulting larvae, and in the relative size of the polar lobe and of the 
blastomeres during the cleavage-stages. Second, a certain pro- 
portion of the entire eggs sooner or later develop abnormally, 
which results in an Increasing mortality from day to day. Third, 
and most important, the percentage of monstrous forms, and the 
mortality. Is always very large in the development of egg-frag- 
ments and of Isolated blastomeres. This Is undoubtedly due in 
part to the abnormal conditions under which the larvae are placed 
In the aquarium, In part to the shock of the operation, and in part 

Experimental Studies on Germinal Localization. 21 

to the changed condition of surface-tension in the dwarf embryos 
and larvae, as is shown by the readiness with which they dis- 
integrate. (I have several times seen an actively SAvimming dwarf 
larva suddenly fly to pieces on coming in contact with an obstacle 
or even with the surface of the water.) For these reasons, de- 
spite the great ease with which the eggs may be operated, it is 
difficult to base trustworthy conclusions regarding the more spe- 
cial features of the egg-localization on the defects observ^ed in 
the individual partial larvae. I have therefore in the following 
work restricted my account in the main to the results that appear 
with unmistakable clearness, and appear in so large a proportion 
of the larvae as to remove all reasonable doubt. Beyond this, 
owing to the importance of following the development of the 
living larvae as far as possible, the number preserved for section- 
ing was not very large, and the technical difficulties indicated 
above, in case of the normal larvae, here appear in aggravated 
form. This explanation is necessary to account for certain ob- 
vious gaps in the work, which I hope to fill out by further in- 
vestigation, especially those relating to the mesoblast, regarding 
which I can at present offer only somewhat provisional conclu- 



(a) General History of the lobeless Larvae. — During the 
trefoil stage of the first cleavage the polar lobe may easily be 
removed, wholly in part, by means of a fine scalpel. Complete 
removal of the lobe produces a highly characteristic and constant, 
though in one respect very unexpected, result. Exactly as Cramp- 
ton earlier found in Ilyanassa, the egg continues to segment after 
this operation quite symmetrically, in a manner similar to the 
normal cleavage of such forms as Patella or Lymnaea, giving rise 
by typically alternating spiral cleavages to successive symmetrical 
quartets of micromeres (Figs. 20-26). These cleavages differ 
constantly in two respects from the normal, namely, that ( i ) 
no trace of a polar lobe is formed at either the second or the 
third cleavage, and (2) the members of the D-quadrant are no 


Edmund B. Wilson. 

larger than the others. Correlated with this is the fact that these 
embryos show no lower white area, all the basal quadrants being 
uniformly pigmented over the lower pole (Figs. 23, 24), which 
sometimes shows a large opening into the cleavage-cavity (Fig. 

Fig. IV. 
Cleavage after Removal of the Polar Lobe. 
20, Two-cell stage and polar lobe after removal of the latter ; 21, four- 
cell stage of same, from upper pole ; 22, eight-cell stage of same, from upper 
pole ; 23, sixteen-cell stage of lobeless embryo from lower pole , symmetrical 
second quartet; 24, similar view of the same stage, open type; 25, sixteen-cell 
stage, from upper pole ; 26, lobeless embryo from lower pole, after formation 
of the third quartet; 27, second cleavage, from the side, of egg from which 
about three- fourths of the first polar lobe had been removed ; 28, a similar 
form, viewed from the lower pole, after removal of about one-half of the first 

Experimental Studies on Germinal Localization. 23 

24) . The embryos gastrulate and develop with great regularity 
into larvae that swim in the same characteristic progressive spiral 
course as that of the normal ones. These larvae (Fig. 32) differ 
from the normal ones in two obvious respects, namely, ( i ) the 
post-trochal region is absent, or represented only by a smoothly 
rounded surface from which no outgrowth takes place, and (2) 
they show no trace of an apical organ. . The first of these results 
fully accords with expectation; for studies in cell-lineage have 
shown, both in annelids and in mollusks, that in forms possessing 
a typical trochophore larva the ectoblast and mesoblast of the post- 
trochal region are mainly derived from the two somatoblasts, 
and I have shown that the first of these cells is certainly and the 
second probably, derived mainly from the polar lobe (or lower 
white area). The second result, on the other hand, is astonish- 
ing, since the region that has been removed is diametrically oppo- 
site to that from which the apical organ develops; but a large 
number of operations have not shown one exception in this re- 
spect and the most convincing corroborative evidence is afforded 
by other experiments presently to be described. 

The structure and subsequent history of these larvae is very 
widely different from that of the normal forms. As the cleavage 
advances the symmetrical cells of the second and third quartets 
close in around the lower pole, frequently followed in greater 
or less degree by the cells of the prototroch; and after the gas- 
trulation this region (the posterior region of the larva) becomes 
somewhat expanded, so that the larva assumes a pyriform shape, 
actvely swimming with the narrower end in front, and rotating 
from right to left like a normal larva. The narrower anterior 
region is uniformly covered with fine vibratile cilia which are 
slightly longer near the anterior pole (as in a normal larva — 
Fig. 32) ; but an examination of more than fifty such larvae 
failed to show a single case in which a true apical tuft was present. 
Sections and total preparations reveal the remarkable additional 
fact that in such larvae, at least in many cases, no apical plate is 
formed, though the lateral areas of proliferation, referred to 
above, are present, as shown in Fig. 40, a, a. In a few cases 
I have found a somewhat vague thickening at the apical pole, 


Edmund B. JVils 


Experimental Studies on Germinal Localization. 25 

Fig. V. 

Normal Mefamorpliis and lobeless Larvae. 

(Excepting Fig. 33 these figures were drawn from living larva, the cilia 
being added from formol preparations and the inner outlines from specimens 
mounted in balsam.) 

29, Normal trochophore of 24 hours (a rather large specimen) ; 30, normal 
trochophore of 32 hours; 31, normal larva of 72 hours, showing foot and shell; 
2)2, larva of 24 hours, after removal of first polar lobe ; 33, vertical section of 
lobeless larva of 24 hours, showing entoblast-plug protruding through the 
blastopore; 34, larva of 72 hours, after removal of first polar lobe; 35I larva of 
24 hours, produced from a form like Fig. 28, after removal of about half the 
polar lobe ; 36, larva of 24 hours, after removal of second polar lobe ; 37, CD 
half-larva, after removal of second polar lobe, 24 hours. 

1 This figure has been turned upside down by the engraver. 

26 Edmund B. Wilson. 

but never one that could be mistaken for a typical apical plate. 
In others, however, the apical ectoderm does not differ from that 
by which the whole pre-trochal region is surrounded. I feel 
justified therefore in the statement that the lobeless larvae typically 
fail to develop the apical organ at any period, individuals having 
been reared up to the fourth day, when the metamorphosis of the 
normal larvae was well advanced. {Cf. Figs. 31 and 34.) Dur- 
ing the development, probably owing to the deficiency of material 
present in the D-quadrant, the trochoblasts often become more or 
less displaced towards the posterior pole, and in greater or less 
degree lose their regular arrangement. In many specimens never- 
theless the typical prototrochal belt of three rows of cilia is formed 
(Fig. 34) , though even in these the rounded posterior region often 
also bears patches of cilia. In others no definite belt can be made 
out, and such individuals often give the appearance, when alive, 
and even after being killed with formol, of being ciliated over the 
whole posterior region. In preparations, however, the cilia of 
such forms may almost always be seen to be arranged in patches, 
leaving non-ciliated regions between them, which are doubtless 
occupied by cells derived from the second and third quartets. 
It is probable, therefore, that the appearance of uniform clllation 
is misleading, and is caused by the confusion of separate tufts 
lying at different levels. In cases where no displacement of the 
trochoblasts occurs, the posterior region is covered by cells de- 
rived from the second and third quartets. 

As the development proceeds there is no attempt to regenerate 
the missing post-trochal region or apical organ, and the later 
history of these larvae differs totally from that of the normal 
ones. The pre-trochal region shows an increase, instead of a 
decrease, in size, and is not withdrawn Into the interior, but gives 
rise to a more or less irregular vesicular structure directed for- 
wards as the embryo swims. Such larvae were reared until the 
beginning of the fourth day (Fig. 34), after which they in- 
variably became more and more irregular and finally disinte- 
grated. At this period they present a most remarkable contrast 
to the normal control larvae of the same age. There is still no 
trace of a post-trochal region, no shell, no foot, and no apical 

Experimental Studies on Germinal Localization. 27 

organ. Sections show that these larvae have formed no shell- 
gland, no mantle-fold, and apparently also no mouth. 

The foregoing account applies to the great majority of the 
lobeless larvae; but occasionally an apparent exception occurs, the 
careful examination of which only serves to confirm the rule. 
In these exceptional cases a more or less reduced post-trochal 
region appears to be present, and one individual was obtained 
that in life seemed to possess this region in a fully developed con- 
dition. Sections of these embryos show, however, that what 
appears to be a post-trochal region is in reality a plug of ento- 
blast cells, projecting through the blastopore-region, that arises 
through defective gastrulation (Fig. 33). Such embryos some- 
times show towards the upper pole a much larger cleavage-cavity 
than in the normal form, — obviously a result of the failure of 
the entoblast-cells to invaginate completely. This is conspicuously 
shown in the larva, referred to above, which appeared to have a 
fully developed post-trochal region. This larva, cut into longi- 
tudinal serial sections, shows very clearly the failure of the ento- 
blast-cells to invaginate properly, a large space being left in the 
upper hemisphere above the archenteron. For this very reason 
this larva showed very clearly, both as a total preparation and 
after sectioning, the entire lack of an apical organ. 

The foregoing observations fully establish the conclusion, I 
believe, that the material of the polar lobe is indispensable for the 
formation of the post-trochal region and the apical organ, and as 
shown beyond they give considerable reason for extending this 
conclusion also to the ccelomesoblast. . That the failure to produce 
a normal larva is not due to the lack of sufficient material, is con- 
clusively shown by several additional facts. First, in Patella 
the D-quadrant is no larger than the others, yet a post-trochal 
region is formed that is relatively as large as in Dentalium. 
Second, as will be described in Part V, much smaller larvae, pos- 
sessing all of the typical parts, may be produced from fertilized 
egg-fragments. Third, the same conclusion is afforded by the 
history of isolated blastomeres, which also fully corroborates the 
results obtained by removing the polar lobe from an entire egg. 
If in the 2-cell stage the two blastomeres, AB and CD, be sep- 

28 Edmund B. Jfilscn. 

arated, both continue to segment for a time as if still forming 
part of an entire embryo, the second and third polar lobes form- 
ing in normal fashion in the CD half; but in the end both com- 
pletely close, gastrulate, and form activ^ely swimming larvae. The 
two larvae agree in possessing a closed, though often somewhat 
asymmetrical or confused prototroch, but otherwise show the fol- 
lowing characteristic and constant differences. The AB (smaller) 
larva, closely resembles, except in size, that derived from an 
entire egg from which the polar lobe has been removed, invaria- 
bly lacking a post-trochal region and apical organ (Fig. 46). 
The CD (larger) larva, on the other hand, possesses both these 
structures, both of which may be as large as in a whole embryo 
(Figs. 42-45). These larvae vary greatly in form, but in gen- 
eral are asymmetrical and, as may be seen by a comparison of 
Figs. 45 and 29, possess a post-trochal region that is almost in- 
variably relatively too large, and a pre-trochal region relatively 
too small as compared with a normal larva. As in the AB half, 
the prototrochal cilia frequently show a confused arrangement, 
the regular rings of the normal larva being more or less broken 
up. In like manner, if the four blastomeres of the 4-cell stage 
be isolated, only the larva from the D (largest) quadrant develops 
these two structures (Fig. 47), while those from A, B or C are 
nearly like those derived from the AB half, though only half as 
large (Figs. 48, 51). Like the CD ^-larvae the D ^^-forms are 
variable in form; but whenever they complete what may be con- 
sidered their normal development they show the post-trochal region 
very much too large, and the pre-trochal region much too small 

(Fig. 47)- 

All these larvae show a very high mortalit}^ but I have kept the 
^-larvae as late as the beginning of the fourth day ( Fig. 51), and 
the 54-larvae nearly as long. The smaller larvae (the AB half, 
or the small quarters) show a greater tenacity of life, swim more 
actively, and become less irregular than the larger ones. In the 
end, however, all the forms become irregular and finally wholly 
disintegrate, without producing normally formed trochophores or 
regenerating the missing structures. The CD ^-larvae of 24 
hours sometimes approach the form of normal larvae of the same 
age, though always showing the false proportions of the pre-tro- 

Experimental Studies on Germinal Localization. 29 

chal and post-trochal regions described above. Like the AB halves 
and the j4-larvae, they often swim actively at the surface, rotating 
in the same way as an entire larva; though the progressive move- 
ment is almost always slower and less regular than that of the 
smaller halves. These forms, however, seem to live no longer than 
the less regular ones, and in spite of every precaution they become 
more and more irregular and finally disintegrate in the same 
aquaria containing the normally developing whole larvae. Those 
that lived to the end of the second day invariably became mon- 
strous in form and showed no resemblance to a normal larva. The 
history of the AB halves or the smaller quarters in general very 
closely resembles that of the lobeless larvae, the pre-trochal 
region enlarging, becoming irregular, and finally disintegrating, 
often while the embryo is still actively swimming by means of 
the trochoblasts, which, as Fischel has observed in case of the 
swimming cells of ctenophores, are most tenacious of life of all 
the cells. 

The relative volumes of protoplasmic substance contained by 
these various forms of larvae, may be determined either by meas- 
uring the volumes of the blastomeres after isolation by means of 
calcium-free sea-water, or by measuring the polar lobe and es- 
timating the other volumes, the two methods giving fairly consist- 
ent results. It should be remembered, howev^er, that both the 
whole eggs and the relative size of the polar lobe (and hence of 
the blastomeres) vary somewhat, both in the eggs produced by 
a single female, and to some extent in those produced by dif- 
ferent females. I observed one. lot of eggs, for instance, the 
greater number of which produced lobes considerably smaller than 
usual. Measurements of the lobe in typical average trefoils give 
a value ranging from one-fifth to one-sixth that of an entire 
egg. A typical case gave a volume of 0.18 for the lobe, from 
which the other volumes are as follows : 

Entire embryo i-oo 

Embryo without polar lobe 0.82 

CD y2 embryo 0.59 

AB y2 embryo 0.41 

D y^ embryo 0.385 

, A, B or C 0.205 

30 Edmund B. Wilson. 

Since the CD >4 larva is less than ^ and the D }i larva less 
than ^ the volume of the lobeless embryo, yet both produce 
apical organ and post-trochal region, the conclusion is unavoid- 
able that the failure to form these structures after removal of 
the polar lobe must be due to a qualitative and not a quantitative 
difference; in other words, the material of the lobe must be spe- 
cifically different from the remaining material, and as such is 
the determining cause of the development of the structures in 

The above conclusion is fully sustained by the effect of cutting 
off only a part of the polar lobe. In such embryos during the 
second and third cleavages the polar lobe is correspondingly 
diminished in size (Figs. 27, 28), and the D-quadrant is too 
small by the same amount. Such eggs produce larvae with a cor- 
responding reduction in the post-trochal region (Fig. 35) and 
these larvae sometimes possess, sometimes lack, the apical organ. 
It is not improbable therefore that further experiments of this 
kind may show a localization, within the polar lobe itself, of the 
determining materials of the apical organ and of the post-trochal 
region. This experiment adds to the foregoing the important 
result that after the polar lobe has formed there is a direct quan- 
titative relation between the amount of specific material it con- 
tains and the size of the post-trochal region, there being appar- 
ently no regulative process in the later stages (though I have not 
yet sufficiently examined this latter point). As will appear in 
Part V, this conclusion does not apply to the material of the lower 
polar area before the formation of the lobe. 

(b) The mesoblast question. — We may now consider what 
is in some respects the most interesting, as it is certainly 
the most difficult, of the questions relating to the lobeless 
larvae, namely, that of the mesoblast. The fact that cer- 
tainly the first and probably the second somatoblast is de- 
rived mainly from the substance of the polar lobe, and 
that after the removal of this substance the post-trochal region 
fails to develop, suggests that the material of the coelomesoblast 
as well as of the ectoblastic structures, is localized in the polar 
lobe and hence in the original polar area. In point of fact 

Experimental Studies on Germinal Localization. 31 

Crampton ('96) in his interesting paper on Ilyanasssa, found 
that after removal of the polar lobe the second somatoblast (4d) 
differs from the normal not only in being no larger than the 
other members of the quartet, but also in texture, being filled 
with yolk-spheres instead of being mainly composed of clear 
protoplasm as in the normal, and it also lies at first at the sur- 
face, exactly like 4a, 4b and 4c. This observation I can confirm 
from a reexamination of the original preparations, kindly placed 
at my disposition for this purpose by Dr. Crampton. He found 
further, that the larvae produced from such eggs lacked the 
mesoblast-bands present in the normal larva, 4d apparently enter- 
ing, like its fellow-members of the same quartet, into the forma- 
tion of the archenteron. 

This highly interesting result, which has atracted considerable 
attention, was based on the examination of total preparations 
only; and the desirability of a more adequate study of the matter 
by means of sections has long been obvious. I have accordingly 
given especial attention to this point as far as my material would 
allow; but must admit that neither in point of abundance nor 
of fixation is this material quite adequate for the full investigation 
of the question, which indeed would demand a complete study of 
the cell-lineage, both in the normal and in the lobeless forms. 
Nevertheless such evidence as I hav^e obtained is distinctly in 
favor of the correctness of Crampton's result. 

The mesoblast may be most clearly seen in the normal larvae 
in cross sections through the region of the prototroch, where the 
gut shows two chambers and the complication produced further 
back by the shell-gland and mantle-folds are not present. In such 
a section (Fig. 38) the gut appears in the form of two distinct 
chambers, the wall of the ventral one being a little further back 
intimately connected with the stomodaeal invagination (Fig. 39') 
though its cavity does not yet appear to communicate with the 
outside. The walls of both chambers are composed of large 
cells, more or less columnar and radially disposed, completely 
filled with yolk-spheres (as are all the cells at this time) and with 
large nuclei. On either side is a loose group of much smaller 
cells with small nuclei, that appear irregular or often spindle- 


Edmund B. Wilson. 

Fig. VI. 

Sections of normal and lobeless Larvae. 

(Each of these is drawn from a single section, supplemented by a few details 
from the two adjacent sections of the series. The deutoplasm is only shown in 
the entoblast and mesoblast.) 

38, Slightly oblique cross-section of normal larva, 24 hours, just anterior to 
the mouth ; 39, cross-section through the mouth ; 40, vertical section of lobeless 
larva, 30 hours; 41, cross-section through prototroch-region of lobeless larva, 
48 hours. 

Experimental Studies on Germinal Localization. 33 

shaped. There can, I think, be no doubt that these are meso- 
blast cells,^ though I have not determined whether they are the 
products of the second somatoblast, 4d, or arise from another 
source. A possibility of error on this point is given by the fact, 
already referred to, that just anterior to the prototroch on either 
side are two lateral ectoblastic areas of proliferation (of unknown 
significance) that may contribute to the small cells in question. 
In any case these lateral masses of mesoblast fail to appear in the 
lobeless embryos of corresponding age or older. . In the earlier 
stages, of which Fig. 33 is an example, it is impossible to de- 
termine this point with any degree of certainty, owing to the 
crowding together of the entoblast cells in a compact mass in 
which frequently no cavity can be seen. In later stages, however, 
both longitudinal and transverse sections give pretty clear evi- 
dence that the small mesoblast-cells are either wholly absent or 
very few in number. Fig. 40 is from a complete series of lon- 
gitudinal sections of a lobeless embryo of 30 hours. This shows 
the gut as a two-chambered sac directly applied to the ectoblast 
with no sign of smaller cells between them, though both the 
anterior ectoblastic areas of proliferation are shown (a, a). It 
might well be supposed that the small cells are present in a dif- 
ferent plane, as would be the case in Fig. 38 if cut in the sagittal 
plane ; but their absence appears no less clearly in cross-section, as 
shown in Fig. 41 (from a complete transverse series). This 
embryo of 48 hours swam actively and normally. Though not 
so well fixed as the preceding one, it clearly shows the gut as a 
simple sac, enclosing a single cavity that opens at the posterior pole 
and anteriorly is nearly filled with a thickening bulging inward 
from the wall at one side. I am quite sure that no mesoblast- 
cells are present in this embryo unless at the extreme anterior end, 
where the layers are cut tangentially and cannot be clearly an- 
alyzed. The sections of this embryo clearly show further 

1 The relations as figured by Kowalewsky ('83, Fig. 48) in the Marseilles 
species are essentially similar to those here shown, except that the mesoblast- 
cells are shown very much larger and fewer. This is stated to be from a larva 
of 24 hours, but probably represents a relatively earlier stage of development 
than mine. Compare the mesoblast-cells in Kowalewsky's Fig. 66, from a larva 
of 38 hours. 

34 Edmund B. JVilson. 

the absence of any structure comparable with the foot, mantle- 
folds, shell-gland, or mouth (unless the posterior opening can be 
so considered) though all these structures are present in the 
normal control embryos. The absence of an apical organ is 
shown as in other series, by the two from which Figs. 33 and 40 
are taken. 

I would not speak too positively before examining additional 
material, for in some of the other series a few small cells appear 
that may be of the same nature as those seen in the normal 
embryos, though they are far less numerous; yet the foregoing 
evidence is sufficient to create a strong presumption that Cramp- 
ton's result was correct. Crampton showed due caution In guard- 
ing against the conclusion, from his observations, that the polar 
lobe "contains prelocallzed mesoblast material," being probably 
Influenced by the fact that in Ilyanassa the lobe appears to be 
composed mainly of deutoplasm. He only concluded "that the 
presence of the yolk mass In the cell D may be the stimulus which 
causes that cell to act differently from the other macromeres, A, 
B and C" ('96, p. 14). I believe, however, the facts brought 
forward In this paper render it probable that the polar lobe (and 
hence the cell D) does in fact contain a specific kind of cytoplasm 
which. If not actually "prelocallzed mesoblast-material" is f.he 
direct and necessary antecedent of that material. 



The failure of the AB half-larva to produce an apical organ, 
though wholly consistent with the history of the lobeless embryos, 
was to me a surprising fact; for the development of this organ 
in other forms Indicates that all of the four quadrants contribute 
to Its formation ; and in point of fact I had found in Patella that 
not only do both the AB and CD halves produce an apical organ, 
but also any of the ^-embryos, and even any Isolated micromere 
of the first quartet. I therefore turned with much Interest to a 
more detailed examination of the localization of this organ in 
Dentalium; and this Involved the inquiry whether the correla- 

Experimental Studies on Germinal Localization. 35 

Fig. VII. 

Larvae from isolated Blastomeres. 

42, 43, 44, Various forms of larvae from isolated CD halves, 24 hours; 45, 46, 
twin larvae from the isolated CD and AB halves of the same egg, 24 hours; 
47, larva from isolated D-quadrant, 24 hours; 48, larva from isolated C-quadrant 
of the same egg, 24 hours; 49, larva from isolated posterior micromere, id, of 
8-cell stage, 24 hours; 50, larva from isolated micromere, ic, of the same egg, 
24 hours; 51, one-fourth larva from one -of the small quadrants (A, B or C), 
72 hours. 

36 Edmund B. Wilson. 

tion between apical organ and post-trochal region is direct or in- 
direct — i. e,, whether the development of the one depends on that 
of the other, or whether the development of the two is only con- 
nected through their common relation to the polar lobe. Further 
experiments conclusively show that the latter is the case; for in 
several ways larvae may be produced that possess the apical organ 
but lack the post-trochal region. My first experiment to test 
this consisted in the isolation, separately, of the four micromeres 
of the first quartet (la, ib, ic, id), which may easily be effected 
by means of Herbst's calcium-free sea-water. The result of this 
experiment, several times repeated, is that while all four of these 
micromeres may develop into actively swimming ectoblastic em- 
bryos, the one derived from the D quadrant ( id) , and this alone, 
develops an apical organ (Figs. 49, 50) . All of these four small 
embryos are of approximately the same size, ovoidal or some- 
what pear-shaped in form, with a group of active trochoblasts at 
the larger (posterior) end. The anterior region is covered with 
fine cilia (as in the AB ^-larya or the A, B or C ^4 -larva) ; 
but only the id larva bears in addition the characteristic apical 
tuft, which is nearly or quite as large as in a whole embryo, and 
is borne upon the usual ectoblastic thickening or apical plate. 
None of these larvae gastrulate or develop a post-trochal region; 
from which it follows that after the completion of the third cleav- 
age not only is the development of the apical organ independent 
of that of the post-trcchal region, but at this time the posterior 
micromere of the first quartet, id, is already definitely specified for 
the formation of that organ, independently of its relation to the 
remainder of the embryo. The result of isolating the cells of the 
4-cell stage is entirely in harmony with this, as already mentioned. 
The A, B or C 34 develops into a closed pyriform larva swimming 
normally with the smaller and turned forwards, but entirely 
devoid of apical organ or post-trochal region (Fig. 48). The 
D ^, on the other hand, though often distorted, shows typically 
the apical organ, and an exaggerated and usually irregular post- 
trochal region. (Fig. 47.) This result is in striking contrast 
to the fact, mentioned above, that in Patella, each of the quadrants, 
whether of the 4-cell stage or of the first quartet, may develop an 

Experimental Studies on Germinal Localization. 37 

apical organ. The only conclusion that can be drawn from this 
contrast is that the definitive basis of the apical organ is more 
closely localized in Dentalium than in Patella, being concentrated 
in a single cell. 

The above results prove that the determination of the develop- 
ment of the apical organ takes place at some period between 
the first and the third cleavages. Further experiments fix the 
period of determination still more nearly. If the egg be allowed 
to advance as far as the second cleavage and the polar lobe formed 
at that time be removed, the egg continues to segment in a manner 
indistinguishable from that of an egg from which the lobe has 
been removed at the time of the first cleavage. From such eggs 
arise larva agreeing exactly with those arising after removal of 
the first polar lobe in every respect save one, namely, that the 
apical organ is typically present, though this is not invariably 
the case. (Fig. 36.) Sections clearly show that the apical 
tuft is borne upon a very definite apical plate, in striking con- 
trast to the larvae arising after removal of the first polar lobe. 
It is thus possible to produce at will larvae which lack the 
post - trochal region and either possess or lack the apical or- 
gan; and the determination of the apical organ is thus proved 
to be effected during the short period between the first and sec- 
ond cleavages. Complete corroboration is given by removal of 
the second polar lobe from the isolated CD ^ during its first 
division. The resulting larva resembles that arising from the 
AB half in having no post - trochal region, but possesses an apical 
organ as well developed as though the polar lobe had not been 
removed. (Fig. 37.) 

The experiments just described prove, first, that the correla- 
tion between post-trochal region and apical organ is due to their 
common determination by the first polar lobe. The second polar 
lobe, though apparently precisely similar to the first, has no longer 
any influence on the apical organ, though it still determines the 
development of the post-trochal region. It seems impossible to 
explain these facts, save under the assumption that the first polar 
lobe contains specific stuffs that are in some manner essential to 
the formation of both structures, and that during the period 

38 Edmund B. Wilson. 

between the first and second cleavages the "apical stuff" (if such 
a term be allowed) exerts once for all its specific effect. The 
most natural explaaation of this is given by the hypothesis that 
this stuff moves upward to the apical pole, to be isolated in the 
large posterior quadrant, D, during the second cleavage, and 
subsequently in the corresponding micromere, id, during the 
third cleavage. The basis of correlation between post-trochal 
region and apical organ may thus be sought in the physical as- 
sociation of the corresponding specific stuffs in the first polar lobe, 
while the specification of the posterior micromere, id, is due to 
the final isolation within it of the "apical stuff." 



The preceding sections are in a measure only preliminary to 
the present one which includes the most important part of the 
present paper, namely, the results of experiments on the localiza- 
tion of the polar lobe, and of the structures that it involves, in 
the unsegmented egg. As has already been stated, the clear sub- 
stance forming the polar lobe is already visible in the egg prior 
not only to cleavage, but even to fertilization and maturation. 
Experiments on the unsegmented egg show with great clearness 
that this area possesses in a general way the same promorpho- 
logical value as the polar lobe itself, though at this early period 
the egg possesses a greater regulative capacity than at later stages. 
The unfertilized living eggs of Dentalium may readily be cut in 
two with the scalpel under the microscope, and the plane of sec- 
tion determined with considerable accuracy not only during the 
operation but by a subsequent examination of the fragments in 
which the polar areas are often still clearly visible. As Yves 
Delage first showed, such fragments when fertilized may segment 
and give rise to ciliated embryos and in certain cases even to 
dwarf trochophores. In a considerable proportion of such ex- 
periments, both fragments develop. For convenience of descrip- 
tion I shall divide them into two classes, including (a) those 
obtained by horizontal or oblique section, and (b) those obtained 

b2 65 

Fig. VIII. 

Development of Egg-fragments after horizontal or oblique ScctionA 

52, 53, Twins, after oblique or horizontal section near upper pole ; 54, trefoil, 
lower half, horizontal section above equator; 55, 56, twins, after oblique section, 
larger lower, smaller upper fragment; 57, 58, 59, twins, after slightly unequal 
oblique section; 57, trefoil from lower fragment, 58, upper fragmertt (failed to 
segment), 59, trochophore of 24 hours developed from 57; 60, 61, 62, twins, 
horizontal section below equator ; 60, trefoil, lower fragment, 61, 2-cell stage 
of same, 62, upper fragment, 2-cell stage; 63, trefoil, lower fragment, horizontal 
section , polar lobe too small ; 63a, trefoil lower fragment, with polar lobe 
slightly too large ; 64, 65, twins, plane uncertain, 64 undeveloped fragment, 65 
trochophore, 24 hours. 

1 In these and the following figures the plane of section is indicated by the 
small accompanying diagram, the fragment studied being marked with a 

40 Edmund B. Wilson. 

by exactly vertical section passing through the axis and bisecting 
the polar areas. 

{a) Fragments obtained by horizontal or oblique section. — 
Under this heading may be grouped all fragments obtained by 
sections passing in such a plane as to separate the polar areas, so 
that one fragment contains only the upper, the other only the 
lower, of these areas. These may be designated respectively as 
the upper and the lower fragments. Before maturation I have 
not found it possible to distinguish the upper from the lower 
fragment; but as soon as the polar bodies form, the upper frag- 
ment may be at once identified with certainty, since it alone pro- 
duces these bodies. I have not thus far observed any difference 
between the results of horizontal and oJ& oblique sections. 

The upper and lower fragments differ in a characteristic way, 
both in the form of cleavage and In the structure of the resulting 
larvae; though it should be added that this appears most clearly 
In the cleavage-process, since many of the embryos die before 
reaching the trochophore stage, and many of the remainder be- 
come wholly monstrous in form. Nevertheless the main result Is 
given with great consistency by a comparison of the larvae. This 
contrast is especially striking when two fragments from the same 
egg are compared; and within rather wide limits It Is Independent 
of the plane of section and the size of the piece, certainly as far 
as the form of cleavage Is concerned, and apparently also as re- 
gards the larval type. Whether large or small the upper frag- 
ment forms the polar bodies In normal fashion, and In many 
cases segments in essentially the same way as an egg from which 
the polar lobe has been removed. The first cleavage takes place 
without the formation of a polar lobe and Is Invariably equal 
(Figs. 53, 56, 62, etc.), and the same applies to the second cleav- 
age. Frequently the two pairs of cells shift during or after the 
second cleavage, so as to produce a "cross-form," the succeeding 
divisions of which are difficult to analyze. In many cases, how- 
ever, the four cells remain In .nearly the same plane ; and In such 
cases the succeeding divisions conform to the regular rule of 
spiral cleavage, quartets of micromeres being found by alternat- 
ing dexlotropic and lelotropic divisions. (Fig. 69.) 

Experimental Studies on Germinal Localization. 41 

A considerable proportion of these embryos fail to develop 
into larvae, breaking up sooner or later Into loose groups of 
cells that perish. Many, however, develop Into actively swim- 
ming larvae, but these, whether large or small, are never normal 
trochophores. While showing many variations, and often being 
more or less Irregular In form, these larvae tend In general 
towards, and sometimes agree precisely with those derived from 
whole eggs minus the polar lobe, from the AB half, or the A, B 
or C quarters (Figs. 70, 86). They are in general more or 
less distinctly pyriform, swimming actively by the long cilia that 
are more or less irregularly disposed about the posterior enlarged 
region. A typical case is shown in Fig. 70 (from a preparation, 
the cilia from the living larva) produced from the upper two- 
thirds of an egg after exactly horizontal section. The cleavage 
of this fragment was similar to that shown In Fig. 69. This 
larva Is In every respect closely similar to the lobeless larva, though 
the pre-trochal region is more expanded than usual, forming 
a large hollow vesicle enclosing a few loose cells, and with a 
slight thickening at the anterior pole, but without anything like 
a true apical organ. The posterior region Is filled with a crowded 
mass of rounded cells. Transverse sections of this larva show that 
this mass Incloses a very small central cavity; but It is Impossible 
to determine whether mesoblast cells are present or not. In a 
very few cases an apical organ Is present in such larvae; but this 
is so rare that I attribute Its occasional presence to the fact that 
the plane of section was not quite correctly determined, a portion 
of the lower polar area having been in fact included in the piece. 
Another possibility is that the specific material of the polar lobe 
extends so far up into the interior as to be removed by a section 
that externally passes quite outside the polar area. This Inter- 
pretation is supported by the fact that In a very few cases, when 
the upper fragment Is considerably larger than the lower one, I 
have seen the upper fragment form a very small polar lobe. 

The development of the lower fragment — i. e., one that In- 
cludes the lower polar area — differs In a remarkable way from 
that of the upper one, both In the form of cleavage and in the 
end-result. Whether obtained by horizontal or oblique sections. 

42 Edmund B. Wilson. 

and (within rather wide limits) whatever its size, this fragment 
may segment in every detail like an entire egg of diminished 
size, forming the polar lobe in normal fashion, and may give rise 
to a dwarf larva nearly or quite normal in form and possessing 
an apical organ. The study of a large number of these fragments 
shows that while there is considerable variation in the size of the 
polar lobe it is as a rule of approximately and often exactly, 
of the correct proportional volume; and this is true eveij after a 
horizontal section that. passes quite outside the limits of the polar 
area. By varying the plane of section it is thus possble to obtain 
a graduated series of forms leading down from a full-sized em- 
bryo to one not more than one-fourth this size, the fragments from 
the other halves forming a similar series grading in the opposite di- 
rection. That the form of cleavage Is within wide limits, inde- 
pendent of the size of the piece, is thus strikingly demonstrated. 
Such a graded series of trefoils and the corresponding equal 
2-cell stages, is shown in Figs. 52-60, the last of these showing 
the smallest one observed. Regulation of the size of the polar 
lobe sometimes fails however, examples being shown in Fig. 63, 
where, even after horizontal section, the lobe is too small (this 
egg produced a larvae possessing an apical organ, but with the 
post-trochal region greatly reduced). Fig. G^a, where it is slightly 
too large, and Fig. 66^ where it is much too large; but these are 
exceptional. It is hardly possible that this apparent regulation 
is owing to the fact that the specific polar material extends so 
far up into the interior of the egg that a section in almost any 
plane includes the right amount of material to form a normally 
proportional lobe. Such an explanation is rendered very improb- 
able by the usual failure of the upper fragment to form a lobe 
even after horizontal section far down in the vegetative hemi- 
sphere or after oblique section; and still more improbable by the 
fact that so many of the fragments form a normally proportioned 
lobe, whatever be the plane of section. The conclusion therefore 
appears unavoidable that the size of the polar lobe, and hence 
of the structures dependent upon it, is subject to a regulative 
process, from which it follows that the predetermination of the 
region of the polar lobe is qualitative, not quantitative, or if 

Experimental Studies on Germinal Localization. 43 

Development of Egg-fragments 
66, 67, Lower fragment, oblique section; 66, trefoil, polar lobe much too large, 
67 resulting 2-cell stage, CD half too large ; 68, 69a,-69c, cleavage of twin frag- 
ments, oblique section, 68 trefoil, from smaller lower fragment, 69a-69c, sym- 
metrical cleavage of larger upper fragment, 4-cell to i6-cell stages; 70, larva 
of 24 hours, from upper fragment of a case exactly similar to 69; 71, 72, twins, 
oblique section, nearly equal fragments, 71, 2-cell stage of upper fragment. 72 and 
72, 2-cell and 8-cell stages of lower fragment; 73, 74, twins, probably oblique 
section near upper pole, 74, upper fragment, 4-cell stage, 73a, normal third cleav- 
age of lower fragment from lower pole, 73b, resulting 8-cell stage, 73c, begin- 
ning of fourth cleavage, formation of first somatoblast. 

44 Edmund B. JVilson. 

quantitative, it is still subject to the operation of a regulative factor 
that lies behind the topographical distribution of the egg-mate- 
rials. This appears to me one of the most significant results that 
my experiments have yielded. 

The embryos may in succeeding stages cleave in every detail 
like whole eggs. Typical 4-cell stages are shown in Figs. 73a, 
78b, 83, 8-cell stages in Figs. 72, 73b, 84, and the fourth cleav- 
age, with the formation of the first somatoblast, in Fig. 73c. Fig. 
73a shows the third cleavage with the formation of the third 
polar lobe. Many individuals were observed showing the forma- 
tion of the second polar lobe in normal fashion, though none are 

The larvae arising from fragments of this type differ as mark- 
edly from those derived from the upper fragments as does the 
cleavage. Although many of the embryos perish, and of those that 
live many are abnormal, they frequently possess both the apical 
organ and a post-trochal region; and occasionally a dwarf larva 
is produced that is essentially similar, except in size, to an entire 
trochophore. One of the best of these is shown in Fig. 59, which 
arose from a lower fragment obtained by oblique section, slightly 
smaller than half the volume of the egg, and including the whole 
of the polar area. The typical trefoil stage of this larva is shown 
in Fig. 57; it has exactly the normal proportions, and segmented 
normally in later stages. This larva Is somewhat less pointed 
posteriorly than the normal, but the whole larvae vary consid- 
erably in this regard. It swam in quite normal fashion. Another 
larger larva from a lower fragment in shown in Fig. 6^. The 
total preparation of this larva shows with great clearness a typ- 
ical apical plate at the upper pole. Out of a very large num- 
ber of operations I have obtained altogether not more than five 
or six such perfect larvae, at least half the embryos dying during 
the cleavage, and a large proportion becoming abnormal during 
the later development. 

That so large a proportion of the embryos die or develop 
abnormally is to be expected when we consider the very different 
mechanical conditions of surface-tension and the like in these small 
embryos. The fact remains that abnormal larvae may be pro- 

Experimental Studies on Germinal Localization. 45 

duced from lower fragments less than half the size of the egg; 
and that such larva may possess a typical apical organ when the 
section passes far away from the apical pole; while in no case does 
the upper fragment produce a larva that ever approaches the 
normal form. It may therefore safely be concluded that the 
dwarf trochophores obtained by Yves Delage ('99) arose from 
fragments including at least a part of the lower polar area. 

The abnormalities observed in larvae from the lower frag- 
ments range from only slight defects to wholly irregular and 
monstrous forms, and thus far do not permit any more detailed 
conclusions regarding the prelocalization than those stated above. 
A common defect, illustrated by the pair of twins shown in Figs, 
85, 86, is a more or less imperfect development of the post-trochal 
region, even when the whole lower area is included in the frag- 
ment, and sometimes this region appears to be wholly lacking. 
Much more rarely the apical organ is lacking while the post- 
trochal region is in greater or less degree developed. Such a 
case is shown in Fig. 87 (from a preparation), the absence of the 
apical tuft having been certainly determined in the living larva. 

As in the case of the lobeless larvae, the experiments dem- 
onstrate that the failure of the upper fragment to produce the 
missing structures is not due to an insufficient mass of proto- 
plasm; for I have obtained larvae showing the characteristic de- 
fects from upper fragments fully two-thirds the bulk of the egg 
(Fig. 70), and perfect dwarfs from much smaller fragments 
(Fig. 59). The conclusion is therefore unavoidable that, like 
the polar lobe to which it gives rise, the lower polar area contains 
specific materials that are essential for the formation of the apical 
organ, and of a post-trochal region; and that it is these materials 
that enter into the formation of the polar lobe, as simple observa- 
tion of the normal development indicates. 

{b) Fragments obtained by ^vertical section through the axis. 
— In view of the foregoing results we should expect to find that 
when the egg is cut exactly vertically, so as to bisect the lower 
polar area, both fragments should form the polar lobe; and such 
is in fact the case. The experiments of this type were not very 
numerous, and only a few cases were obtained In which both frag- 

46 Edmund B. Wilson. 

ments developed. I have only one pair of camera sketches to 
show the polar lobes in such a case (Fig. 75,76). In both these 
the lobe is relatively too small, as if produced from insufficient 
material; but this not always the case (as shown beyond), and 
it should be remembered that the polar lobe is sometimes too 
small even in a lower fragment containing the whole of the lower 
polar area (Fig. 62,). Figs. 77a, 78a show a pair, one of which 
has a lobe of normal proportions ; the »ther is a very nearly nor- 
mally formed 2-cell stage, though the larger cell is perhaps a trifle 
too small. Both these produced nearly normally proportioned 4- 
cell stages (Figs. 77b, 78b). Several other cases, in which only 
one fragment developed, showed a normal trefoil. These data are 
somewhat meagre, yet they justify the conclusion, I believe, that 
after vertical section bisecting the lower polar area both frag- 
ments may segment like whole eggs of half size. 

The above conclusion renders it probable that by such vertical 
section two perfect dwarf trochophores may be produced from 
a single egg, which is apparently impossible when one fragment 
alone contains the lower polar area. In point of fact, I have 
never obtained even a single wholly normal larva after such sec- 
tion; but in view of the comparatively small number of successful 
operations and the very small number of such larvae obtained 
by section in other planes this is not surprising. A number of 
larvae from more or less nearly vertical sections is shown in the 
following figures. Fig. 88 is a nearly normally formed larva 
with two apical organs, from an oblique section passing outside 
the lower white area. Fig. 89 is a nearly normal larva from a 
section that removed a part of the lower area. Fig. 93 is from 
an exactly vertical section bisecting both areas. In section this 
larva Is closely similar to a normal one, and seems to show that 
the trochoblasts are as large as in a whole embryo. Fig. 90 Is 
from the smaller fragment after a slightly oblique section bi- 
secting the lower area ; a very distinct apical organ is present and 
also an abnormally formed post-trochal region. Figs. 91, 92 are 
twins from a slightly unequal vertical section (developed from the 
respective twin fragments 81, 82), the post-trochal region is 
lacking In both, while one lacks an apical organ. 

Experimental Studies on Germinal Localization. 47 

Fig. X. 

Development of Egg-fragments after vertical Section. 

75, y6, Equal twins, respectively in trefoil and polar lobe-formation; lobes 
too small ; 77, 78, equal twins, nearly correct proportions ; 77a, 77b, typical trefoil 
and 4-cell stages of one fragment; 78a, 78b ,typical 2-cell, slightly abnormal 
4-cell stages of the twin fragment; 79, 80, twins, from a fertilized egg, 79, 
nearly normal trefoil, 80, the twin, with reduced polar lobe; 81, 82, nearly 
equal twins. 81, typical 4-cell stage, 82, its twin, nearly typical 2-cell stage; 
83, 84, typical 4-cell and 8-cell stages, from upper pole, of the same fragment 

48 Edmund B. Wilson. 

It may be pointed out that not one of these larvae shows a 
fully developed post - trochal region, though 91 and 92 arose 
respectively from 2- and 4-cell stages that show nearly the normal 
proportions and must have been produced from nearly normal 
trefoils. This may seem to contradict the conclusion, drawn 
above, that the predetermination of the lower polar area is not 
quantitative ; but a similar reduction sometimes exists in this region 
when the whole polar area is present (as in Fig. 85), and I do 
not think a trustworthy conclusion can be drawn without addi- 
tional data. 

I may add that after a large number of unsuccessful attempts 
I obtained two nearly normal dwarf trochophores from frag- 
ments of the unsegmented egg of Patella. One of- these, which is 
about half the volume of a normal larva, clearly shows the cells 
of the prototroch. In the full - sized normal trochophore of 
Patella the prototroch, as may be seen with the greatest clearness 
in total preparations, consists of a closed principal ring of cells 
that vary in number (as seen in optical section) from 19 to 21. 
In the dwarf the cells are more variable in size and less regularly 
arranged, but on the average as large as in the normal individual; 
equatorial optical section of this larva shows 13 cells in the prin- 
cipal ring. 



Extremely interesting and curious results are obtained by a 
comparison of the behavior of fragments of fertilized eggs, and of 
the isolated polar lobe, with that of fragments of the unfertilized 
eggs described above. 

{a) The behavior of fragments of fertilized eggs obtained 
before cleavage. — In order to malce sure that the eggs were fer- 
tilized the operation was delayed until one or both polar bodies 
had been formed, and the egg was then cut as nearly as possible 
horizontally, so as to separate the lower polar area from the 
nucleated part. As already described, if this operation be per- 
formed on the unfertilized egg, and the two fragments be fer- 
tilized, both may, and frequently do, develop. When, however, 

Experimental Studies on Germinal Localization. 49 

Fig. XI. 
Larvae of 24 Hours, from Egg-fragments. 

85, 86, Twin larva of 24 hours, oblique section passing outside lower polar 
area, 85, the lower, 86, the npper larva; 87, larva from lower two-thirds, hori- 
zontal section, without apical organ; 88, larva from lower two-thirds, oblique 
section, two apical organs; 89, larva from nearly vertical section; 90, larva 
from smaller fragment, slightly oblique section bisecting lower area; 91, 92, 
twin larvae, produced from 81 and 82 respectively, vertical section; 93, larva 
from exactly vertical section. 


Edmund B. Wilson. 

Experimental Studies on Germinal Localization. 51 

Fragments of fertilised Eggs, horizontal Section; isolated Polar Lob'es and/ 

Fragments of Lobes. 

94, 95, Equal twins from the same egg ; 9Sa, 95b, upper half, 2- and 8-cell 
stages ; 943-941, successive changes in the lower enucleated fragment ; 94a, soon 
after operation; 94b, first polar lobe (drawn immediately before 9Sa) ; 94c, first 
resting stage, upper fragment in 2-cell stage (23 m. after b) ; 94d, second polar 
lobe (21 m. after c, the upper fragment just divided into 4) ; 94e, second rest 
(16 m. after d) ; 94f, third polar lobe (i6m. after last, at nearly .the same time 
with 9Sb) ; 94g, third rest (7 m. after last) ; 94h, fourth lobe (46 m after last, 
fourth cleavage in progress in upper fragment) ; after 16 minutes the' fragment 
appeared to be divided into two and so remained ; 94i, the same four hours 
later; 96, lower fragment, like last, but showing correct proportions of first 
polar lobe; 97, successive changes in isolated polar lobe from the individual 
shown in Fig. 21; 97a, soon after removal; 97b, first active period (16 m., the 
egg just divided into 4) ; 97c, ensuing first resting period (14m. after b) ; in the 
second period of activity (not sketched), 15 m. later, as the eggs divided into 
8, the lobe constricted as in b, but not so deeply, and again became spherical in 
a second resting period ; 97d, e, third period of activity, 74 and 78 ni. after the 
first period; 97f, final result, 36 m. later; 98, another isolated lobe, 98a, third 
resting period; 98b, second lobe; 98c, final result, in which condition it re- 
mained without further change; 99, 100, two fragments obtained by cutting 
a polar lobe in two (the original lobe was slightly larger than usual), showing 
active changes shortly after division of the egg into 4; 99a, looa, are shown 
42 m. after 99 and 100; 99b, loob, 8 m. later; 99 remained in this condition, 
while 100 again became spherical by fusion of the two halves and underwent 
no further change. 

52 Edmund B. JVilson. 

a fertilized egg is thus sectioned only the nucleated {i. e., the 
upper) fragment develops — a result that agrees with my ob- 
servations on the nemertine egg and that of Renilla ('03), and 
with the earlier ones of Delage ('01) on those of echinoderms. 
This fragment has essentially the same mode of development 
as a corresponding fragment of an unfertilized egg, segmenting 
equally into two and four without the formation of a polar lobe, 
forming successive symmetrical quartets of micromeres by alter- 
nating spiral cleavages (Fig, 95), and producing a larva that is 
either an irregular monster or a pyriform larva closely similar 
to those arising from the lobeless egg or the AB half. This is 
what would be expected in view of the preceding results; but 
the behavior of the non-nucleated lower half is most remarkable 
in that it forms three times in succession a polar lobe from the 
white area at the same time that the nucleated half is dividing^ 
becoming spherical after each period of activity without dividing. 
When this was first observed, I believed that I must in some way 
have confused the fragments with those of unfertilized eggs; but 
repetitions of the experiment under conditions that precluded all 
error, gave the same result. A typical case is shown in Fig. 94, 
from consecutive camera drawings of the same fragment. The 
first lobe is shown (Fig. 94b ) about 15 minutes after the opera- 
tion, while the nucleated half (Fig. 95) has just divided into 
equal halves. Twenty-three minutes later the fragment was again 
perfectly spherical (94c), while the upper fragment was in a 
resting 2-cell stage. The second lobe (94d) was formed 44 
minutes after the first, while the upper fragment was dividing into 
4 equal cells, after which the lower fragment again became spher- 
ical (94e, 16 minutes later than 94d). The third lobe (94f) 
was formed 32 minutes after the second, and was considerably 
smaller than either the first or the second, as in a whole egg; 
the upper fragment meanwhile divided into eight cells (Fig. 95b) . 
A third period of rest followed (Fig. 94g) . Following the fourth 
cleavage of the upper fragment the lower one passed through a 
change no less remarkable than the preceding (it is at this 
period in the normal development that a large part of the 
lower polar area passes into the first somatoblast) . This 

Experimental Studies on Germinal Localization. 53 

change begins with the formation of a fourth lobe, com- 
posed of white material, which is at first much smaller than 
any of the preceding (94h, 46 minutes after 94g). Unlike the 
preceding lobes this one was not resorbed into the fragment, but 
was permanent, slowly increasing in size until after two or three 
hours it was nearly as large as the remaining portions, the frag- 
ment now appearing as if divided into two (94i). 

This case is fairly typical of several that were followed through 
the entire cycle of changes, and one or more of the stages were 
seen in many individuals. The lobes are not always so distinctly 
formed as in the one figured, and the final stage, though usually 
like that described, varies considerably in appearance. 

{b) Behavior of the isolated polar lobe. — Previous to mak- 
ing the observations just described, I had several .times observed 
changes of form in the isolated polar lobes after their removal 
from the trefoil stage. On reexamining the matter I found that 
these changes are also periodic, taking place approximately at the 
same time as the cleavage in the lobeless nucleated portion. The ac- 
tivities of the isolated lobe at these periods vary considerably in 
different individuals. Sometimes the activity is no more than a 
slight change of form, the spherical lobe becoming slightly pyri- 
form or even almost amoeboid. Frequently, however, the isolated 
lobe actually forms a smaller lobe by a process that closely sim- 
ulates the formation of a polar lobe by a whole egg or an egg- 
fragment. In any case, each period of activity is followed by a 
spherical resting-stage that coincides approximately in time with 
the resting stages of the segmenting lobeless portion. I regret 
that I had not time to study this remarkable phenomenon with 
sufl[icient care, but give series of sketches illustrating two particular 
cases. Fig. 97a shows a lobe soon after its removal; 97b, the 
same, 16 minutes later just after the egg had divided into four; 
97c, the ensuing resting stage, 14 minutes after 97b; a second 
period of activity followed, in which the lobe again constricted, 
but not so deeply as at 97b, followed by a second spherical stage; 
97d and 97e show the third active period, and 97f the final result, 
after which no further change occurred. In 98 is shown the final 
active period of a lobe, which resulted in the permanent apparent 

54 Edmund B. Wilson. 

division of the lobe into two. Even if the lobe be cut in two 
after its removal, the fragments likewise pass through alternating 
periods of activity and rest closely similar to those of the whole 
lobe, as is shown in Figs. 99, 100 (the original lobe was somewhat 
larger than in the other cases shown). This proves that the 
power of a rhythmic change of form involving the temporary 
formation of lobe-like structures, is not a property of the lobe as 
a whole, or of the lower polar area, but is inherent in the sub- 
stance of which it is composed. It would be interesting to com- 
pare in this respect the behavior of the isolated lobe, or fragment 
of a lobe, with fragments from other regions of the fertilized egg. 
Such fragments would probably also exhibit rhythmic changes, but 
I hazard the conjecture that their activity would be found to 
differ in some definite way from that of the lobe-fragments. 

The phenomena above described, which deserve further careful 
study, are of interest both cytologically and embryologically. 
First, since both the nuclei and the centrosomes are absent, it 
follows with great probability that even in the cleavage of a 
whole egg the constriction of the cell that leads to the formation 
of the polar lobe takes place wholly independently of either these 
structures or the astral rays, which suggests the possibility that 
the same may be true of the constrictions that lead to complete 
cell-division. Second, since the rhythm in the formation of the 
polar lobes in the enucleated fragment coincides with that shown 
in the division of the nucleated fragment, it is clear that as far 
as the lobe-formation is concerned the cytoplasmic division rhythm 
is quite independent of that of either the centrosome or the chro- 
mosomes. This fact may be placed behind the one earlier de- 
termined by Boveri ('97), Zlegler ('98) and myself ('01), that 
the rhythmic activities of the chromosomes and of the cen- 
trosomes are likewise independent, or at least separable. But 
beyond this it Is remarkable that the periodic activity in the non- 
nucleated fragment is not merely of a rhythmic character, but 
changes its character at the time of the fourth cleavage when in 
the normal development the material of the polar lobe no longer 
forms a merely temporary structure, but is permanently cut off 
by a cell-division. We here catch a glimpse, as It were, of a 

Experimental Studies on Germinal Localization. 55 

definite order of events predetermined in a particular cytoplasmic 
area and wholly independent of the immediate action of nucleus 
or centrosome. An additional point of great embryological in- 
terest is the fact, shown by a comparison of Fig, 6 with Fig. 94, 
that in these fragments the polar lobe is, at least in some cases, 
nearly or quite as large absolutely as in one entire egg; whereas 
in the lower fragment of an unfertilized egg it is typically re- 
duced to the correct proportional volume of the lobe in a whole 
egg. This is however not invariable, for in some cases, an ex- 
ample of which is shown in Fig. 96, the lobe is reduced to its 
proper proportional size. I have not accurately studied this mat- 
ter in a sufficient number of cases to speak very positively; yet I 
feel confident that the contrast in this respect between the lower 
fragments from unfertilized and fertilized eggs is a general, 
though not an invariable rule. The interest of this fact is 
pointed out in the sequel. 



Without undertaking at this time a complete discussion of the 
foregoing observations, I may briefly indicate their bearing on 
the general questions referred to at the beginning.^ My observa- 
tions demonstrate conclusively, I think, both the mosaic character 
of cleavage in these eggs, and the definite prelocalization of some 
of the most important morphogenic factors in the unsegmented 
egg. The Dentalium egg shows, even before it breaks loose from 
its attachment in the ovary,and long before even the initial changes 
of maturation, a visible definite topographical grouping of the 
cytoplasmic materials. This is proved by the experiments to 
stand in definite causal relation to the subsequent differentiation 
of the embryo in such wise that the removal of a particular cyto- 
plasmic area of the unsegmented egg results in definite defects 
in the resulting embryo that are not restored by regenerative or 
other regulative processes within the time-limits of the experi- 
ment. Since both the egg-fragments and the isolated blastomeres 

1 A more general discussion of the mosaic-theory of development, with a 
fuller review of the literature, will be given in a following paper. 

56 Edmund B. fVilson. 

become perfectly spherical before development proceeds, the re- 
sulting defects cannot be due to a failure of regulation traceable 
to the shape of the fragment, as was formerly assumed by- several 
writers. Neither are they due to insufficient mass; for perfect 
dwarfs may arise from fragments much smaller than those that 
show the characteristic defects. Further, these facts, like those 
earlier determined by Crampton ('96) in the gasteropod egg, and 
by Driesch and Morgan ('95) and more recently by Fischel 
('98) in the ctenophore egg, are fatal to the view that embryonic 
differentiation is brought about through quahtative nuclear divi- 
sion during the cleavage. The conclusion is therefore unavoidable 
that the specification of the blastomeres in these eggs is due to 
their reception, not of a particular kind of chromatin, but of a par- 
ticular kind of cytoplasm; and that the unsegmented egg con- 
tains such different kinds of cytoplasm in a definite topographical 
arrangement. How many such specific stuffs exist in the unseg- 
mented egg of Dentalhim and what is their arrangement it is 
impossible at present to say; for the pigment-band and the two 
polar areas can only be considered as an outward sign of an or- 
ganization that for the most part doubtless escapes the eye. My 
experiments have only positively determined the cytoplasmic pre- 
localization in the lower polar area of material essential for the 
development of that complex of structures that I have included 
in the term "post-trochal region," and of one other structure, the 
apical organ. The first of these includes material that is essential 
to the development of the typical larval form, including the foot, 
to certain characteristic ectoblastic structures of the post-trochal 
region, such as the shell-gland, mantle-fold, and probably also the 
pedal ganglia ; it also appears probable that it includes material 
essential for the formation of the coelomesoblast. I do not doubt 
that further experiments on this egg will show a still more definite 
and detailed prelocalization ; though, as already stated, it is 
not easy to determine this, owing to the difficulty of distinguishing 
between defects in the partial larvae that result directly from the 
plane of section and those that are due to other causes. 

Two additional facts clearly appear from the experiments, on 
which I would lay stress. First, the amount of material removed 

Experimental Studies on Germinal Localization. 57 

with the polar lobe or lower polar area is wholly disproportionate 
to the effect produced. The polar lobe includes less than one- 
fifth the volume of the egg; yet its removal does not merely cause 
a structural defect of like extent, but inhibits the whole process 
of growth and differentiation in the post-trochal region and the 
concomitant withdrawal of the pre-trochal region. The cleavage 
of the lobeless embryos shows that both the second and the third 
quartets are formed; and it is fair to conclude that certainly in 
the AB half of the embryo, and probably also in the CD half, 
these cells contain ectoblastic material, which in a normal embryo 
would contribute to the formation of the post-trochal region. 
These cells, as stated above, close in around the posterior region, 
and perhaps are partially turned in with the invaginating ento- 
blast-cells. In any case, however, the power of active growth in 
the post-trochal region, so conspicuous in the normal larva, is 
wholly lost with the removal of the excess of material in the D 
quadrant. It does not seem possible that this loss in power of 
growth is due to mechanical obstacles, since the same defects exist 
in fragments of the unsegmented egg from which the lower 
polar area has been removed and which are free to segment as 
best they can. The conclusion therefore appears unavoidable that 
the material of the lobe is not only specifically necessary for the 
formation of the bases of the post-trochal structures, but also for 
the whole growth-process that is here brought to a focus. Apart 
from its more general bearings, this conclusion is important from 
the light that it may throw on the teloblastic growth of annelids 
and other segmented forms, and it seems altogether probable 
that if the polar lobe could be removed from such an egg as that 
of Sabellaria or Myzostoma the resulting larva would fail to 
develop a metameric trunk-region. 

A second point of interest that clearly appears from the ex- 
periments is that the topographical grouping of specific materials 
in the unsegmented egg may be in its ensemble widely different 
from that of the definitive bases of the organs which they de- 
termine; for the experiments demonstrate that the apical organ, 
lying at the upper pole, is determined by material originally lying 
far down in the vegetative hemisphere in the lower polar area. 

58 Edmund B. Wilson. 

On this point an analogous result has recently been obtained by 
Yatsu, who has shown with great probability that in the unseg- 
mented nemertine egg the basis of the apical organ does not lie 
at the upper pole, where we should expect to find it, but in, or 
slightly above, the equatorial region. 

These facts have an important bearing on our interpretation of 
development in general. In my previous paper on the nemertine 
egg I have developed an hypothesis of differentiation agreeing 
broadly with Sach's well-known theory of formative stuffs, and 
with the general conclusions regarding mosaic development inde- 
pendently published by Fischel ('03) nearly at the same time, the 
essential assumptions being that the prospective value of a cell is de- 
termined by its cyVoplasmic content, that this content is de- 
termined by the form of cleavage in connection with an antecedent 
formation and segregation of specifically different materials 
(which may Itself determine the form of cleavage), and that 
the morphogenic function of cleavage, so to say, is to isolate 
the materials thus segregated. This conception, it is hardly neces- 
sary to point out, receives very definite support by the observations 
now brought forward; but I wish to bring them more closely into 
relation with those made on the nemertine and echinoderm eggs, 
especially with regard to the general question of progressive (i. e., 
epigenetic) localization in the egg. In the nemertine {Cerebra- 
tulus) I found that either an isolated blastomere or a fragment 
from any region of the unsegmented egg may produce a perfect 
dwarf larva; but the two differ In the form of cleavage, the 
blastomere segmenting as if still forming part of a whole em- 
bryo and producing an open blastula (as in the echinoderm), 
while the egg-fragment segments like a whole egg and produces 
a closed blastula — that is, it develops as a whole from the be- 
ginning. I explained the contrast In development between the 
two as the result of a regrouping of the egg-materials, occurring 
during and subsequent to the process of maturation and fertiliza- 
tion, which Initiates the morphogenic process and determines also 
the form of the earlier cleavages. I pointed out that such re- 
grouping of materials is known to occur at the maturation-period 
of many eggs — for Instance, in the sea-urchin — and suggested 

Experimental Studies on Germinal Localization. 59 

that the contrast between the development of an egg-fragment in 
the nemertine and in a sea-urchin (where it segments like a whole 
egg only after section in certain planes) is owing to the fact that 
in the latter, egg-fragments have only been obtained in the period 
subsequent to maturation when the regrouping has been effected. 
Localization of the cleavage-factors was thus conceived, essen- 
tially in agreement with Roux's early conclusions regarding the 
frog's egg, as a progressive (i. e., epigenetic) process, and the 
same conception was applied to the general morphogenic process 
which, as is shown with especial clearness by the facts here brought 
forward, may be so closely connected with the cleavage-process. 

As far as the progressive character of localization is con- 
cerned, the result obtained in Dentalium may seem at first sight 
to be in disagreement with the conclusions just reviewed, for the 
germ-regions are here defined by a definite segregation of ma- 
terials that exists even in the attached ovarian egg long before 
either maturation or fertilization, and the isolated blastomere 
is not capable of producing a complete embryo. But the contra- 
diction disappears upon comparison with certain other forms, 
which are intermediate in character between the extremes repre- 
sented by Dentalium and the nemertine or echinoderm egg; and 
this comparison demonstrates, as I believe, the validity of the 
theory of "precocious segregation," formulated as a pure specula- 
tion by Ray Lankester in 1877. I have already expressed the 
opinion that the horizontal stratification of the egg expressed by 
the three zones of material visible in Dentalium or Myzostoma is 
comparable, or at least analogous, to that which finds an expres- 
sion in the formation of the well-known polar rings of leeches and 
oligochaetes. This comparison is based both on the position and 
mode of formation of these rings and on their fate. Vejdovsky 
('88) very clearly shows that in Rhynchelmis both the polar rings 
arise as local thickenings of a general ectoplasmic layer, and both 
assume at one period the form of protoplasmic discs lying at 
either pole of the egg (as Whitman also observed in Clepsine) . 
Except for the fact that the upper and lower protoplasmic areas 
have not at any period been seen to appear in the form of actual 

6o Edmund B. JVilson. 

rings, the resemblance to these relations of those observed in 
Dentaliiim is unmistakably obvious. It is entirely possible that 
the correspondence is not complete; but that in a general way 
the resemblance indicates a similar form of stratification in the 
molluscan and annelidan egg, seems hardly open to question; and 
the comparison is sustained by the fact that in Clepsine both rings 
were traced by Whitman into the AB half, and the upper one 
into the D quadrant, while in Rhyfichelviis Vejdovsky traced both 
rings into the D quadrant, where the material of the two fuses into 
one mass in the 4-cell stage and later passes into the mesomeres, 
which are undoubtedly to be identified with the somatoblasts.^ 
If this comparison be admitted a further comparison of these 
and some other forms is highly significant. In Dentaliiim three 
structural zones are present from the beginning, the lower one 
coinciding in extent with the lower white area, the upper one 
lying at the centre of the upper white area, at first very small, but 
rapidly increasing in extent during and after the maturation 
period. A condition similar to this exists in Sternaspis, where 
Vejdovsky ('81) showed that a distinct protoplasmic area, which 
he compares to a polar ring ('88, p. 122) lies at each pole of 
the ovarian egg, the upper one being much smaller than the lower 
one, though larger than in Dentaliiim. . In Clepsine and Rhyn- 
chelmis three structural zones are likewise present, but tJiese first 
appear during the maturation period with the development of the 
polar rings, like the three zones described by Boveri ('01) in the 
Strongylocentrotus egg. The egg of Myzostoma occupies, at 
least in some respects, an intermediate position. No upper pro- 
toplasmic disc has here been observed as yet, but the lower proto- 
plasmic area is obviously represented by the green mass, which, as 
Driesch ('96) has shown passes into the polar lobe, and subse- 
quently certainly in part into the first somatoblast, and probably 
in part into the second somatoblast, precisely as in Dentaliiim. 
The interest of this case, compared with the foregoing, lies in the 
fact observed by Driesch (which I can confirm) that before ma- 

1 Cf. Vejdovsky and Mrazek ('03, p. 454) ; see also the highly interesting 
statement (p. 534) that the dense protoplasm of the polar rings ("Polplasmen") 
can be recognized as such "in den Zellen des Mesoblasts inbesondere in den 
grossen Mesomeren." 

Experimental Studies on Germinal Localization. 6 1 

turation the egg shows at first but two colored zones, of which 
the lower green one exactly represents the lower white area of 
Dentalium, while the upper one first segregates during maturation 
into an upper red zone and an equatorial colorless one. Like 
the lower zone the two upper ones correspond very closely in fate 
to those in Dentalium; for the upper (red) area passes into the ec- 
tomeres, like the upper white area of Dentalium, while the middle 
(colorless) zone passes into the entomeres, as is the case with the 
greater part of the middle (pigmented) zone in Dentalium. It is 
possible that sufficiently careful search may reveal the presence 
in Myzostoma of an upper protoplasmic disc, comparable with 
a polar ring; and as far as the visible colored zones are concerned, 
it is evident that the Myzostoma egg stands midway between those 
of Dentalium and Strongylocentrotus , and it is probably inter- 
mediate also between Dentalium or Sternaspis and Clepsine or 

It seems a legitimate interpretation of the foregoing series 
that these eggs present an essentially similar form of stratification 
which is attained at different periods in the ontogeny, and that 
as compared with the leech or oligochaete, Myzostoma and Den- 
talium or Sternaspis represent two earlier stages in the precocious 
segregation of specific cytoplasmic materials that have a like pros- 
pective value in the development.^ But if this be admitted, it 
follows that in none of these cases can the segregation in question 
be considered as a primary character or "preformed quality" of 
the^egg. .Upon this secondary localization of material, as my 
experiments prove, depend many of the most important features 
of the later morphogenic localization; and I think a presumption 
Is thus established that cytoplasmic prelocallzation is in general of 
like secondary or epigenetic origin, though to what extent this 
holds true can only be determined by further experiment. 

Although the characteristic segregation is In its main outlines 
effected very early In the egg of Dentalium, It may be pointed out 
that, like so many other eggs, there is the clearest evidence of 

1 Cf. Vejdovsky "Wahrend aber bei Sternaspis die Concentration des Bil- 
dungsplasma an beiden Polen bereits im Laufe der Eibildung stattfindet, sam- 
melt sich dasselbe bei Rhynchelmis erst nach der Polzellenbildung und dem Ein- 
dringen des Spermatozoon in das Ei an" ('88, p. 122.) 

62 Edmund B. JVilson. 

later movements and progressive segregation of the cytoplasmic 
materials. I will only call attention, among these, first, to the 
determination of the apical organ by material originally lying In 
the lower polar area, which, if my interpretation of the experi- 
ments is valid, moves upwards to the apical pole in the period 
between the first and second cleavages. That such a movement 
occurs is only a matter of inference ; but this interpretation appears 
to me far simpler and more intelligible than to assume a brief 
"Fernwirkung," or the like emanating from the first but not the 
second polar lobe. It Is however not a matter of inference but 
of fact that the remaining material of the lower white area moves 
upwards and towards one side in the 8-cell stage preceding the 
fourth cleavage, when it apparently fuses with the material of the 
upper white area in the D-quadrant. It Is interesting to com- 
pare this with the facts described by Vejdovsky in Rhynchelmis, 
where the remains of the upper and lower polar rings fuse in the 
D-quadrant at the 4-cell stage. 

I have endeavored to show that cytoplasmic prelocalizatlon 
in Dentalium differs only In degree from the conditions existing in 
such eggs as those of the nemertine or sea-urchin. The same may 
be said, I think, of the development of isolated blastomeres, de- 
spite the fact that in Dentalium such blastomeres are incapable 
of producing complete dwarf embryos. As In the nemertine or 
sea-urchin, although the isolated blastomere segments as a part 
and not as a whole, the embryo finally closes, in the course of 
which process structures like the prototroch, the post-trochal and 
pre-trochal regions, and the gut, close to form whole structures. 
That this process, which in the case of the nemertine I compared 
to Morgan's "morphallaxis" In regenerating planarians or hy- 
drolds, falls short of producing a complete embryo in Dentalium, 
may be due to different causes in different cells. In the AB half 
or one of the smaller quarters this Is obviously due in the main 
to lack of the specific material of the lower polar area. The 
failure of the CD half or the D quarter may In part be due to 
a like cause; but since these embryos contain the materials (those 
contained in the lower polar area) that are missing In the other 
cases, their failure may be due to a different cause. The CD half 

Experimental Studies on Germinal Localization. 63 

larvae are sometimes nearly normally formed except for the false 
proportions of the post-trochal and pre-trochal regions. Their In- 
variable subsequent degeneration Into Irregular and monstrous 
forms Is not Improbably due to the abnormal mechanical condi- 
tions created by their mode of development. It seems possible, 
however, that if these larvae could sustain themselves sufficiently 
long they might in some cases succeed in attaining a normal con- 
dition. They die before attaining this end; and hence succeed no 
better than the AB halves In the "attempt" to produce a perfect 

One cause of the difference between the isolated blastomeres of 
the nemertlne or sea-urchin and the mollusk thus doubtless lies 
in a difference In the segregation-pattern such that In the former 
the specific materials are symmetrically divided between the first 
two blastomeres, while in Dentalium such is not the case. In the 
former, accordingly, the earlier cleavages are purely quantitative, 
but in the latter are qualitative as far as the cytoplasm is con- 
cerned, and to this extent produce from the first cleavage onward 
a mosaic-work in entire accordance with Roux's general concep- 
tion, as I long since indicated in the case of Nereis ('94). But 
beyond this the results especially of Driesch's later studies on the 
isolated blastomeres of sea-urchins indicate that here, although 
a definite polarized segregation of material has taken place at the 
time of the earlier cleavages (directly proved by Boveri's ('01) 
observations on Strongylocentrotus, indirectly by Driecsh's ('00) 
comparison of the development of the upper and lower quartets of 
the 8-cell stage) this segregation Is not only symmetrical with 
respect to the axis but is also less definite or less complete than 
In the molluscan egg, — again a difference which finds its nat- 
ural explanation In the theory of precocious segregation (or dif- 
ferentiation). I should therefore interpret the differences be- 
tween the isolated blastomeres of the mollusk and those of the 
sea-urchin or nemertlne as due to a difference, on the one hand. 
In the pattern, on the other hand in the degree, of segregation. 

It is hardly necessary to point out that the foregoing conclusions 
will in large measure reconcile the apparent conflict between the 
fact of cytoplasmic prelocallzation and the continually Increasing 

64 Edmund B. Jf^ilson. 

evidence that the primary determining factors of development 
are to be sought in the nuclear organization. The well-known 
hybridization experiments of Boveri ('92, p. 469) and Driesch 
('98) on sea-urchins have shown that the earlier cleavage-factors 
conform to the maternal type and hence must be predetermined 
in the egg-cytoplasm; and up to the blastula-stage, at least, the 
embryos remain of the pure maternal type. But the same ex- 
periments demonstrate no less clearly that the nucleus begins to 
affect the cytoplasmic phenomena at least as early as the late 
(prismatic) gastrula, and according to Boveri's latest work ('03) 
as early as the mesenchyme-formation, though the latter point is 
disputed by Driesch ('03). It therefore appears possible, not 
to say probable, that every cytoplasmic differentiation, whether 
manifested earlier or later, has been determined by a process in 
which the nucleus is directly concerned, and that the regional 
specifications of the egg-substance are all essentially of secondary 

Another question, which has been often discussed, is raised 
by these observations, namely, as to the relation in the regenerative 
process between the moulding of the mass as a whole (which 
falls under the general conception of Roux's "Umordung der 
Zellen" or Morgan "morphallaxis) and the specification of the 
individual cells. Like the facts determined by Fischel ('98) in 
the ctenophore egg (following the earlier work of Driesch and 
Morgan) those observed in Dentalium bring out with great clear- 
ness the independence, in this case, of the two groups of factors 
by which these are determined. It is a very noteworthy fact that 
all the partial larvae that lack the lower polar area, whatever 
their size or mode of origin, tend to assume the same form, and 
all are alike devoid of further regenerative capacity. The larvae 
arising from entire eggs after removal of the polar lobe only, 
the CD half from which the second polar lobe has been removed, 
the AB half, the A, B or C quarter, or an upper fragment, of 
any size, of the unsegmented egg — all these typically assume 
the characteristic pyriform shape with the trochoblasts surround- 
ing the larger posterior end. This form, which results after 
closure of the embryos and gastrulation, is essentially a prolate 

Experimental Studies on Germinal Localization. 6^ 

spheroid modified by the presence at one end of the large tro- 
choblasts which have not like the other cells the power of con- 
tinued multiplication, and it evidently represents a state of equilib- 
rium towards which any segmented mass of the egg tends that 
is devoid of the lower polar area. Whether the closure of the 
embryos (which in the case of isolated blastomeres are at first 
strictly partial structures) to produce this form should be con- 
sidered as a regulation or regenerative process is largely a ques- 
tion of definition.^ In any case the facts very clearly show that 
the process is not perceptibly influenced by the nature of the cells 
individually considered; nor does it, on the other hand, appear to 
exert any appreciable effect on the nature of the individual cells 
("Umdifferenzierung" of Roux), as will be more clearly shown 
in my second paper.- Certainly the closing of the embryos does 
not lead to the least perceptible tendency towards the restoration 
of the missing structures that are dependent on the material of 
the lower polar area.^ I am in agreement with the opinion of 
Fischel ('98) that, whether a regulative process or not, the 
closing in to form a closed structure is probably explicable as a 
result of relatively simple physical factors, though I doubt whether 
the explanation is as simple as Fischel assumes in the case of the 
ctenophore."* It is difficult to avoid the conclusion that these 
same factors are operative in the establishment of the normal 
form in a whole embryo; but to them are add in the material 
of the lower polar area a far more complex group of factors, at 
present not analyzable, that involve the whole process of growth 
and metamorphosis. That a mass of cytoplasm so small should 

1 Roux ('93, p. 837) interpreted the closure of the open blastula as part of the 
regenerative process, in opposition to Driesch ('92, p. 585), who asserted that 
this had nothing to do with the regenerative process proper ; though he after- 
ward took the ground that it should be considered as an initial regulative pro- 
cess ('96, p. 88). Morgan ('01, p. 13, etc.) classes morphallaxis under the 
head of regeneration, though not the closing in of a cut surface, which is con- 
sidered as a preliminary process. Cf. Child, on "Mechanical Regulation" ('02). 

2 Cf. Crampton, '97, p. 55. 

3 Cf. the remark of Driesch, based especially on Crampton's experiments on 
Ilyanassa; "1st, wie bei Gastropoden und Anneliden, echte Lokalisation der Bil- 
dungsfaktoren im Ei anztmehmen, so schliesst das eine Regulation zum Ganzen 
wirklich aus." ('96, p. 89.) 

■* Cf. Rhumbler, '02, Zur Strassen, '03. 

66 Edmund B. Wilson. 

exert so great an effect on the morphogenic process is a most 
convincing piece of evidence in favor of the theory of specific 
formative stuffs in development. The only intelligible view of 
the polar lobe seems to me to be that it is, so to say, a reservoir 
of such stuffs destined for allotment to particular cells which 
thereby become definitely specified, irrespective of their subsequent 
relation to the embryo as a whole. This is a very different result 
from the oft-quoted one of O. Hertwig that the lineage of par- 
ticular structures from particular blastomeres is nothing more 
than an incidental result of the continuity of development. It is 
equally opposed to the conclusions of other writers who have too 
hastily rejected the principle of mosaic development for which 
Roux and others have contended. 

Lastly I may point out that in so far as these observations show 
the course of differentiation, and the correlation of parts, to be 
determined by a preexisting topographical grouping of specific 
egg-materials they sustain an essentially mechanistic (as opposed 
to a vitalistic) interpretation of development. To conclude how- 
ever that these eggs are devoid of regulative capacity would be 
to overlook some of the most striking of the phenomena I have 
described. The experiments give clear evidence that a power of 
regulation exists in the unsegmented egg that is no less striking 
in form, if more limited in degree, than in the nemertine or echlno- 
derm. As in the case of the nemertine, the typical spiral cleavage, 
alternately dexiotropic and lelotropic, is not affected by section in 
any plane. Far more striking is the fact that in the cleavage of an 
egg-fragment the size of the polar lobe, on which the proportions 
of the trochophore largely depend, is proportional to the size of 
the piece. Since this Is true even after horizontal section, when the 
whole of the lower polar area is included in the piece, it follows that 
the predetermination of this area is qualitative, but not quantita- 
tive, or only quantitative in so far as it is subject to regulative 
control by other factors. This conclusion receives further sup- 
port from the one reached above that the material of the lower 
polar area Is as such specifically concerned not merely with the 
formation of the structures that arise from it but with the form 
of growth that results In the metamorphosis. But if this par- 

Experimental Studies on Germinal Localization. 67 

ticular area shows such a qualitative, as distinguished from a quan- 
titative, pre-determination, one is led to suspect that a like con- 
clusion may apply to other egg-regions, such as those that form 
the gut, the prototroch, and the like ; and to conclude that however 
detailed a prelocalization may exist in the form of regional seg- 
regations of material, a regulative factor may always be present 
that controls their normal combination. In this respect the un- 
segmented egg, may, I believe, be directly compared with such 
an adult animal as a planarian or hydroid, which, while possess- 
ing more or less definitely specified tissues. In a typical grouping, 
nevertheless may possess a high regulative capacity shown in the 
process of regeneration after injury. 

The facts observed give as little clue to the nature of the regu- 
lative factors by which the quantitative relations are determined 
in the egg-fragment as in the fragment of a planarian or hydroid; 
but one or two considerations deserve brief mention. It is note- 
worthy that although the polar lobe regularly forms in the non- 
nucleated vegetative half of a fertilized egg it is as a rule, though 
not always, not reduced, but nearly or quite as large as in a whole 
egg, whereas in a fertilized fragment, representing the same re- 
gion of an unfertilized egg, the lobe is as a rule reduced to its 
proper proportional volume. While I would not lay too much 
stress on this without further study, it seems to indicate that the 
power of regulation, on which the size of the lobe depends, is 
more complete in a nucleated fragment than in an enucleated one. 
Second, when once the polar lobe has formed, the power of regu- 
lation seems to be lost, at least temporarily; for if a part of it 
be cut away the second lobe is of correspondingly reduced size, 
as is also the post-trochal region of the resulting larva. This 
result is supported by the fact that, like the post-trochal region 
to which it gives rise, the polar lobe in the first (virtual second) 
division of the isolated CD half, though sometimes slightly re- 
duced, is in general nearly or quite as large as in a whole embryo. 
These facts prove that the size of the lobe is not determined 
merely by the size of the piece, but by more complex conditions 
existing apparently for only a brief period, and apparently also 
more effective in a nucleated than in a non-nucleated protoplasmic 

68 Edmund B. PFilson. 

mass. This sufficiently Indicates the complexity of the problem 
with which we are dealing, and the importance of further more 
precise studies of the facts. At the same time, it seems clear that 
the problem of proportionate development in a fragment of an 
organism here appears in a much simpler form than In a blastula- 
fragment, or a piece of an adult organism such as a planarlan or 
a hydra ; and I think we should not abandon the hope of finding 
for It a relatively simple solution. While I am not able to offer 
such a solution, It seems to me that it would be rash to deny its 
possibility, not merely in the present instance, but In all analogous 
processes, even when they take place under the more complex 
conditions existing In multicellular masses. 



1. The Dentalium egg shows from the beginning three hori- 
zontal zones, an equatorial pigment-zone and two white polar 
areas. Each of the polar areas includes a specially modified pro- 
toplasmic area probably comparable to a polar ring. 

2. During cleavage the pigmented zone Is allotted mainly to 
the entomeres, the upper white area to the ectomeres, the lower 
white area to the first and probably also the second somatoblast. 
At the first, second and third cleavages the lower white area tem- 
porarily passes into the "yolk-lobe" or polar lobe. 

3. Removal of the first polar lobe leads to a symmetrical 
cleavage without the subsequent formation of polar lobes, and to 
the formation of a larva devoid of post-trochal region and apical 
organ. Removal of a portion of the first lobe produces a larva 
with reduced post-trochal reglon,^ and with or without apical organ. 
Removal of the second polar lobe produces a larva without post- 
trochal region but with an apical organ. 

4. The lobeless larvae undergo no metamorphosis, form no 
foot, shell-gland or shell, no mantle-folds, no pedal ganglia, ap- 
parently no mouth, and probably no coelomesoblast-bands. 

5. The isolated AB half or A, B, or C quarter, produces a 
closed larva closely similar except In size, to the lobeless ones. 
The isolated CD half or D quarter produces a larva possessing a 

Experimental Studies on Germinal Localization. 69 

post-trochal region as large as In a normal larva, and an apical 
organ, which dies without undergoing metamorphosis. The CD 
half from which the second polar lobe is removed produces a larva 
like that from an AB half, but possesses an apical organ. 

6. The isolated micromere id produces a mass of ectoblast- 
cells bearing an apical organ, while la, ib, ic produce no apical 

7. Fertilized fragments of the unsegmented unfertilized egg, 
obtained by horizontal or oblique section, differ in development 
according as they do or do not contain the lower white area. 
The upper fragment segments symmetrically without the forma- 
tion of polar lobes and produces a larva similar to the lobeless 
ones. The lower one segments like a whole egg of diminished 
size, and may produce a normally formed dwarf trochophore. 

8. Fragments obtained by vertical section through the lower 
white area may segment like whole eggs and may produce nearly 
normally formed dwarf trochophores. 

9. Enucleated fragments, containing the lower white area, 
of fertilized eggs, pass through alternating periods of activity 
and quiescence corresponding with the division-rhythm of the 
nucleated half, and form the polar lobes as if still forming part 
of a complete embryo. The same is true of the isolated polar 

10. The foregoing observations demonstrate the prelocallza- 
tlon of specific cytoplasmic stuffs In the unsegmented egg and 
their isolation in the early blastomeres. The lower white area 
contains such stuffs that are essential to the formation of the 
apical organ and the complex of structures forming the post- 
trochal region, including the shell-gland and shell, the foot, the 
mantle-folds and probably the coelomesoblast. These stuffs are 
contained in the first polar lobe, but the second lobe no longer 
contains those necessary for the basis of the apical organ. Pro- 
gressive changes therefore occur in the original distribution of 
the specific cytoplasmic materials. 

11. Comparison Indicates that the conditions observed In the 
molluscan egg differ only In degree from those In the nemertlne 
or echlnoderm. These differences reduce themselves to differ- 

70 Edmund B. PFilson. 

ences in the period of segregation (or differentiation) and in its 
pattern, and are explicable under the general theory of precocious 

12. The early development of egg-fragments indicates that 
the specification of the cytoplasmic regions is primarily qualitative, 
but not quantitative, or if quantitative is still subject to a regulative 
process that lies behind the original topographical grouping of 
the egg-materials. 

13. The development of the molluscan egg is in its essential 
features a mosaic-work and sustains the theory of "Organbiln 
dende Keimbezirke." 


BovERi, Th., '92. — Befruchtung : Merkel u. Bonnet, Ergebn. I. '9i-'92. 

'97. — Zur Physiologic der Kern- und Zelltheilung: Sitzungsber. d. 

phys.-med. Ges. Wiirzburg. 
'01, I. — tjber die Polaritat des Seeigeleies : Verb, phys.-med. Ges. Wiirz- 
burg. (N. F.) XXXIV. 
'01, 2. — Die Polaritat von Ovocyte, Ei und Larve des Strongylocen- 

trotus lividus: Zool. Jahrb. (Anat.-Ontog.) XIV, 4. 
'02. — tJber mehrpolige Mitosen. Verb, phys-med. Ges. Wiirzburg. (N. 

F.) XXXV. 
'03. — liber den Einfluss der Samenzelle auf die Larvencharaktere der 
Echiniden : Arch. Entwm., XVI, 2. 
Child ,C. M., '02. — Fission and Regulation in Stenostoma : Arch. Entwm.,X V,2. 
CoNKLiN, E. G., '99. — Protoplasmic Movements as a Factor in Differentiation : 
Wood's Hole Biol. Lectures. 1898. 
'02. — Karyokinesis and Cytokinesis in the Maturation, Fertilization and 
Cleavage of Crepidula and other Gasteropoda : Journ. Acad. Nat. 
Sci., Phila. II. Ser. XII, i. 
Crampton, H. E., '96. — Experimental Studies on Gasreropod Development ; Arch. 
Entwm. III. 

'97-— The Ascidian Half-Embryo : Ann. N. Y. Acad. Sci., X. • 
Delage, Yves, '99. — fitudes sur la merogonie : Arch. Exp, Zool. (Ser. III.), VII. 
'01. — fitudes experimentale sur la Maturation cytoplasmique chez les 
Echinodermes: Arch. Exp. Zool. (Ser. Ill) IX. 
Driesch, H., '92. — Entwickelungsmechanisches : Anat. Anz., VII, 18. 

'96. — Betrachtungen iiber die Organisation des Eies und ihre Genese : 

Arch. Entwm. IV. 
'98. — tlber rein -miitterliche Charaktere an Bastardlarven von Echiniden : 

Arch. Entwm., VII. 
'99. — Resultate und Probleme der Entwickelungsphysiologie der Thiere : 

Merkel u. Bonnet, Ergebn. VIII. 
'00. — Die isolirten Blastomeren des Echinidenkeimes : Arch. Entwm. X, 

Experimental Studies on Germinal Localization. 71 

2, 3. 

'02. — Neue Antworten und neue Fragen : Ergebnisse, Merkel u. Bon- 
net, XI. 

'03. — Ueber Seeigelbastarde : Arch. Entwm. XVI, 4. 
Driesch and Morgan^ '95. — Zur Analysis der ersten Entwickelungsstudien des 

Ctenophoreneies: Arch. Entwm. II. 
FiscHEL, A., '97. — Experimentelle Untersuchungen am Ctenophorenei, I : Arch. 
Entwm. VI, i. 

'98.— Id., II-IV; Ibid. VII., 4. 

'03. — Entwickelung und Organ-Differenzirung : Ibid., XV. 
KowALEWSKY, A., '83. — fitude sur I'embryogenie du Dentale : Ann. Mus. d'Hist. 

Nat. de Marseille. Zool. I, 7. 
Lacaze, Duthiers, '57. — Histoire de I'organisation et du developpement du Den- 
tale : Ann. Sci. Nat. IV. Ser. VI, VII. 
Lankester, E. Ray, 'jy. — Notes on Embryology and Classification : Q. J. M. S. 

LiLLiE, F. R., '01. — The Organization of the Egg of Unio; Journ. Morph. 

XVII, 2. 
Morgan, T. H., 'oi. — Regeneration: Columbia Biological Series. VIII. 
Rhumbler, L., '02. — Zur Mechanik des Gastrulationsvorganges, etc. : Arch. 

Entwm., XIV. 
Roux, W., '85. — tJber die Bestimmung der Hauptrichtungen des Froschembryo 

im Ei, etc. : Gesammelte Abhandlungen, II. 20. 

'88. — tJber die kiinstliche Hervorbringung "halber" Embryonen, etc. : 
Ges. Abh., II. 22. 

'93. — Ueber Mosaikarbeit und neuere Entwickelungshypothesen : Ges. 
Abh., 27. 

'95. — Gesammelte Abhandlungen. II. 2>3- Nachwort. 

'95. — Uber die verschiedene Entwickelung isolirter erster Blastomeren: 
Arch. Entwm. I. 

'03. Ueber die Ursachen der Bestimmung der Hauptrichtungen des 

Embryo im Froschei : Anat. Anz., XXIII, 4-7. 
Vejdovsky, F., '81. — Untersuchungen iiber die Anatomic, Physiologic und Ent- 
wickelung von Sternaspis : Denkschr. d. Akad. Wien, XLIII. 

'88-'92. — Entwickelungsgeschichtliche Untersuchungen : Prag. 
Vejdovsky und Mrazek, '03. — Umbildung des Cytoplasma wahrend der Be- 

fruchtung und Zellteilung: Arch. Mik. Anat. LXII, 3. 
Wheeler, W. M., '97. — The Maturation, Fecundation and Early Cleavage of 

Myzostoma glabrum : Arch. Biol. XV. 
W HITMAN, C. O., '78.— The Embryology of Clepsine : Q. J. M. S., XVIII. 
Wilson, E. B., '92. — The Cell-lineage of Nereis : Journ. Morph. VI. 

'93. — Amphioxus and the Mosaic Theory of Development: Journ. Morph. 

'94. — The Mosaic Theory of Development : Wood's Hole Biol. Loct. 
Lect. II. 1893. 

'g6, I. — On Cleavage and Mosaic-work: Arch. Entwm. III. 

'96, 2.— The Cell: ist Ed. New York. 

'01. — Experimental Studies in Cytology, II : Archiv. Entwm. XIII., 3. 

72 Edmund B. Wilson. 

'03. — Experiments on Cleavage and Localization in the Nemertine-egg: 
Arch. Entwm. XVI. 3. 

'03.— ^Notes on Merogony and Regeneration in Renilla : Biol. Bull. IV. 5. 
Yatsu, N., '04. — Experiments on the Development of Egg-fragments in Cere- 

bralatus: Biol. Bull., VI, 3. 
ZiEGJ-ER, H. E., '98. — Experimentelle Studien iiber die Zelltheilung, I : Arch. 

Entwm. VI. 2. 

'98. Die Furchungszellen von Beroe ovata : Ardi. Entwm. VII. 

ZuR Strassen, O., '03. — Ueber die Mechanik der Epithelbildung; Verb. d. 

Deutsch. Zool. Ges., 1903. 




With 6 Figures. 


The several experiments, of which this paper presents a resume, 
were conducted during the early summer of 1903, at the Naples 
Zoological Station, while occupying the table of the Smithsonian 
Institution, for the courtesy of which it is a pleasure to express 
my obligations. 

The primary object of the experiments was to test the regen- 
erative capacity of the Scyphomedusae and to institute certain 
comparisons between these results and those obtained by similar 
experiments previously made upon the Hydromedusae. So far 
as I am aware no similar experiments have been made upon the 
Scypliomedusae with the definite purpose of testing this particu- 
lar aspect of their physiological constitution. Romanes in his 
experiments upon " Primitive Nervous Systems," '85, has record- 
ed incidentally the fact that certain mutilations of medusae are 
promptly healed, but gave no details. Eimer, '78, has also 
carried on similar experiments and with the same general purpose 
of testing the character and distribution of nervous centers, but 
makes no reference to the matter of regeneration. And quite 
recently Uexkiill, '00, has likewise reviewed these experiments 
of Romanes and Eimer and carried them somewhat farther than 
they had done. But while arriving at somewhat different con- 
clusions, drawn from a series of experiments in some features 
coincident with those to be described now, he makes no reference 
to any regenerative processes, devoting attention almost exclusively 
to the movements, specially those of rhythmic character, and seek- 
ing physical explanations of them. 

74 Charles W . Hargitt. 

The earlier references of Haeckel to the capacity of larvae of 
certain medusae to regenerate entire organisms are likewise in- 
definite. Morgan in referring to the subject in his recent book on 
"Regeneration," 'oi, merely remarks that among Scyphozoa 'the 
jelly-fishes belonging to this group have a limited amount of re- 
generative power." 

I very much regret that an unusual scarcity of material compels 
me to leave several points somewhat less fully considered than is 
desirable, but I trust they are not of sufficient gravity to seriously 
mar the general value of the results as a whole. 

In one respect this scarcity of material, making necessary suc- 
cessive experiments on the same specimen in many cases, proved 
fortunate rather than otherwise, since facts of importance were 
thus brought to light which might otherwise have been overlooked. 
Some of these will be referred to specifically in another connec- 


The experiments were performed upon Rhizostoma pulmo, one 
of the most common of the Mediterranean medusae. Both in 
size and vigor this medusae affords one of the most satisfactory 
forms for experimentation which has come under my observation. 
It seems likewise to suffer less under the somewhat artificial con- 
ditions of the aquarium than any other which I have had occasion 
to use. As compared with Aurelia and Cyanea of New England 
waters it is incomparably superior in every way, but particularly 
in its ability to thrive for weeks in an environment which would 
prove fatal to the others in as many days. With the single excep- 
tion of Gonionemus I know of no other medusa which affords 
so good a type for this sort of observation and experimentation. 
It was not unusual to have specimens under direct observation in 
the ordinary aquaria of the laboratory rooms for from four to 
six weeks and without apparent deterioration, even in some cases 
under the severe tax of extensive mutilation made necessary by 
the experiments to which they were subjected. It should be stated 
however that as a rule younger and smaller specimens proved 
much better than those of larger size; the latter, on account of 

Regeneration in Rhizostoma Puhno. 75 

their greater mass, are inclined in most cases to sink toward the 
bottom of the tanks, where after a time certain disorganizing 
influences appeared to set up pathologic conditions which seemed 
to deplete their vigor and at the same time render their regen- 
erative processes less satisfactory. 

The experiments were directed to three ends, namely to deter- 
mine: I, The capacity of the medusae to reproduce lost parts, or 
to recover from such Injuries as might ordinarily happen to them 
in a state of nature, such as the battering effects of waves, the 
injuries inflicted by enemies, etc. 

2, The comparative powers of the various regions to regenerate, 
or In other words, the relation of the regenerative capacity to 
liability to Injury. 

3, The capacity to regenerate such highly specialized organs as 
rhopalla, or other sensory structures. 

The experiments included specimens of sizes from about 20 
m/m to 125 m/m In diameter, and while all proved to have 
unexpected powers of regeneration those of medium size, from 
40 to 70 m/m, proved very much more satisfactory than those of 
larger size both in convenience and In their promptness In re- 
sponding to the several sorts of operations, and they apparently 
were more healthy and vigorous during the progress of the ex- 
periments than were those of larger size. Those having a size 
of 100 m/m or more In diameter proved to be much less prompt 
In regeneration and, as will be seen in the records of experiments, 
were much more liable to deteriorate or utterly collapse than were 
the smaller specimens. This is only what might be more or less 
expected, and Is quite In keeping with observations on other 
classes of organisms. The same tendency was more or less evi- 
dent in specimens on exhibition in the public aquarium in which 
of course no mutilations or similar Injuries had occurred. In this 
connection may be noted a somewhat anomalous pathological phe- 
nomenon observed in large specimens both in the exhibition aqua- 
ria and in the small aquaria during the course of experlmentaton, 
namely, the appearance of whitish blotches, or patches of disin- 
tegrating tissues at various places on the exumbrella of the animal 
which sooner or later affected its health and general behavior. 

76 Charles W . Hargitt. 

The matter will be referred to in further detail in another con- 
nection and some reference made as to its probable significance 
and cause. 

In all cases the primary experiments were made as soon as 
possible after the medusae were brought into the laboratory. I 
have said the primary experiments. This refers to the fact al- 
ready alluded to, that in several cases experiments were variously 
repeated upon the same specimen. This was in part for the pur- 
pose of testing the conclusiveness of preceding experiments, and 
in part owing to the fact that there was an insufficient supply of 
material to serve the demands of the course of experiments under 
way. Details as to these aspects will be given in connection with 
the several experiments described. 

The first experiment was made upon a large specimen, and in 
order to determine at the outset whether the earlier observations 
of Romanes and others, that complete removal of the marginal 
sense organs resulted in complete paralysis of the medusa, these 
organs were carefully removed by means of triangular incisions 
as indicated in Figure i, a. The results were substantially con- 
firmatory of the earlier records, the medusa becoming more or 
less passive, except for an occasional single contraction at very 
irregular intervals. This experiment was made on May ii, and 
the following series of observations will suffice to show the general 
course of events. It should be added in this connection that along 
with the excision of the rhopalia several other marginal excisions 
were made, and that three of the oral arms were cut off close 
below the region of the gastric enlargement. The aspect of the 
specimen on the next day was practically the same. While there 
was an occasional contraction of the bell accompanied by certain 
'movements of the body, there were no Indications of rhythm. 

May 13th. — The medusa, while apparently in perfect health 
and vigor of general functions, was still unable to originate any 
definitely rhythmic movements, though responding to various 
mechanical stimuli, such as a strong current of water from the 
tap, or the touch of a glass rod. At various times during the 
day there was evident a rather marked tendency toward sponta- 
neous movements, and occasionally something very like a rhythm, 

Regeneration in Rhizostoma Pulmo. 


several contractions following each other in regular succession, 
though never continuing beyond three or four pulsations. 

May 14th. — The medusa, while still more or less passive as 
before, was yet apparently recovering more of the power of spon- 
taneity, several pulsations occurring at more frequent intervals, 
but these were not of sufficient vigor to produce any locomotion. 


Fig. I. 

Diagram showing methods of excising rhopalia. 

a, usual triangular excision; ai, excision of larger mass; b, rectangular form 
of excision; c, circular form of excision; d, form of rhopalium and lappetts. 

May 15th. — During this and the following day there was an 
apparent relapse of the medusa to the condition of the first day. 
There was also less vigor apparent, such stimuli as those referred 
to above producing but slight effects. This condition continued 
during the i8th, 19th and 20th. 

7 8 Charles JV. Hargitt. 

May 2 1 St. — The medusa seemed to have recovered the vigor 
or tone to which reference has been made above. There was 
also a very evident rhythm in the contractions, often as many as 
ten or more regularly recurring pulsations occurring at irregular 
intervals during the day. As before, however, they were not of 
sufficient force to secure the locomotion of the animal. The same 
condition was observable during the following day. 

May 23d. — There was again a marked decline in both vigor 
and general tone of the body, which showed evident signs of de- 
generation. This condition continued during the following day, 
and on the morning of the 25th the medusa was found to have 
died during the preceding night. 

Upon careful examination it was found that wherever tissue 
had been mutilated or excised there had been a definite healing 
of the wounds and in the case of the oral arms there were indica- 
tions of new growth. I was not able to distinguish that there had 
been any regeneration of the sensory organs, and this will appear 
somewhat surprising in the light of the following experiments. 
Whether there had really been no regeneration at all, or that I 
had overlooked the new organs, or whether they may have dis- 
integrated during the night following the death of the medusa I 
am unable to say. Certain it is, however, that if regeneration 
had gone forward as markedly as in the following cases one could 
hardly have failed to distinguish it. I am inclined to believe 
that the paralysis following the total removal of these organs 
may have served to delay or inhibit active regeneration. 

The next series of experiments differed materially from the 
former, particularly in that care was taken to retain certain of 
the rhopalia in order to insure continued activity of the organisms 
during the progress of the experiment. The number of rhopalia 
retained varied from one to eight, the latter case serving as a 
means of testing the relative influence of these bodies on the 
behavior of the animals and the rate of regeneration. 

On May 12th several specimens, averaging only about half the 
size of the preceding, namely, about 50 m/m in diameter, were 
experimented upon. In the first one all the rhopalia were retained, 
but marginal notches were made of varying sizes between the 

Regeneration in Rhizostoma Pulmo. 79 

sensory bodies, and several of the oral arms were excised. In 
other specimens a varying number of the rhopalia were excised, 
and in one case all the oral arms were cut off close to the gastric 
enlargement and on one side including a portion of this organ 

I shall not undertake to transcribe in detail the records of each 
day, but give rather summaries of results as briefly as is com- 
patible with clearness, trusting that nothing of importance may 
be sacrificed in the attempt to bring the records within as brief 
compass as possible. 

One of the first effects distinguishable in these and following 
experiments was the evident quickening of the pulsations of the 
medusae by the process of excision of the organs, or similar op- 
eration. Not only was the rate of the rhythm greatly increased, 
passing from about seventy pulsations per minute as an average 
for medusae of this size, to ninety, or even one hundred per 
minute. And this rate continued during the entire day, or at 
every observation, which was quite frequent, and well on into 
the second day, when the rate fell to ninety and later to eighty; 
but it was not till the third day that the rate had fallen to the 
normal of seventy per minute. An examination at this time 
showed an evident healing of the wounds and some signs of re- 
generation. Had this been restricted to the sensory bodies it 
might have been interpreted as signifying some important rela- 
tion of these organs to rhythmic activity, but the fact that similar 
effects were produced upon specimens which had not been de- 
prived of their rhopalia would sufficiently negative such an in- 

Eimer, '74, had noted such an effect following a division of 
medusae, particularly those which had been divided into halves 
or fourths, and had undertaken to show that it was chiefly an 
expression of the reduced size of the organism due to its division, 
citing the normal rhythm of specimens of varying size as strongly 
suggesting such an inference. 

Romanes, '85, however, was not able to confirm Eimer's con- 
tention either in reference to matter of fact or the cause assigned. 
Romanes, while citing the variation as to the rate of rhythm in 

8o Charles IF. Hargitt. 

specimens of similar size, is inclined to emphasize what he terms 
the prepotent influence of certain of the lithocysts (rhopalia) in 
coordinating the rate of movement, and the presence or absence of 
such prepotent organs in the portions of medusae under exam- 

Forbes, '48, had long previous called attention to the fact of 
these quickened movements under the influence of various stimuli, 
citing particularly a result of an experiment which he had made 
of a similar character to those which I have cited above. In an 
experiment in which he had, as he expresses it, "paralyzed one 
half of the animal" by cutting out the rhopalia from one side, he 
finds "that the other half contracted as usual, though with more 
rapidity, as if the animal were alarmed or suffering." He remarks 
farther that "all medusae when irritated become much more rapid 
in their movements and contract or expand their disks or bodies In 
a hurried and irregular manner, as if endeavoring to escape from 
their persecutors." (Naked Eyed Medusae, p. 3.) 

While in certain details the conclusions of Forbes may be ques- 
tioned, of his general observations as to matters of fact there 
can hardly be doubt. Furthermore, whether the suggestions of 
either Eimer or Romanes are more than approximate guesses, 
the later observations of Uexkiill have rendered doubtful. So far 
as my own experiments have gone they hardly touch the problem 
of the cause of such reactions. We may safely conclude that, in 
any case, they are of the nature of responses to any continued 
physical stimulus, such as the experiments under consideration cer- 
tainly were. With the healing of the wounds there would of 
course ensue a decline of the irritation, which in turn would be 
followed by a return to the normal rate of rhythm. 

On May 26th, or two weeks following the operation, the me- 
dusae had measurably regenerated all the excised organs. The 
notches cut In the umbrella margins had grown out to complete 
the normal symmetry and there had been developed in the areas 
the characteristic purple pigment, differing from the color of the 
uninjured portions only in its Intensity. The new rhopalia were 
apparently normal in everything save size and pigmentation. 

It is rather noteworthy that in these experiments certain of the 

Regeneration in Rhizostoma Piilmo. 8i 

organs which among the Hydromedusae are most promptly re- 
generated are here among the most slow to develop ; such, for 
example, as the oral arms and gastric lobes. The fact that in the 
rhizostomous medusae these organs have no very active function 
in the capture of food might apparently afford some plausible 
reason for this difference in the rate of regeneration. In Goni- 
onemus the gastric and oral organs are among the most prompt 
in regeneration, and are, of course, also among the most important 
in the functional activities of the animal. That this, rather than 
liability to injury, should be a predisposing factor in regeneration 
would seem to be confirmed in the case of Rhizostoma, for as will 
appear in later experiments there seems to be no good reason to 
suppose that the liability to injury, to which these organs are 
constantly exposed, has anything to do with the capacity for rapid 
or perfect regeneration. 

Additional experiments were begun on May 28th and 30th. In 
this series the specimens varied in size from 20 to 60 m/m in 
diameter. As remarked above there was in these cases the same 
degree of promptness in the responses, which was markedly in 
contrast with that shown by specimens of considerably larger 
size, but in the present cases there was also apparent a somewhat 
less favorable response In the very small specimens. This fact 
considered in connection with the difficulty of operating easily 
upon small specimens, emphasizes the value of animals of medium 
size for such experiments. This conclusion was emphasized 
throughout the entire course of experimentation. 

In part of the specimens of this series only three rhopalia were 
excised, in others four, in others five. In some the rhopalia were 
all removed from one side, while in others only alternate organs 
were removed. In some specimens the same order was observed 
as to excision of mouth arms and other similar operations. One 
of the specimens of the series had only one full-sized mouth arm, 
while the others were in what seemed to be various stages of 
regeneration. As is well known these organs among medusae of 
this type are among the most open to accident from attack of 
fishes or other predatory enemy. The specimen under considera- 
tion would seem to confirm the results of these experiments that 

82 Charles W. Hargitt. ' » 

these organs are readily regenerated, and that in a state of nature 
as well as under the artificial conditions of the laboratory. An 
examination made with the hand lens on June 2d, or only four 
or five days following the operation, showed the first indication 
of regenerating rhopalia. As the organ first makes its appearance 
it is a very minute papilla-like body, and in these cases at the 
inner, or upper edge of the notch made by the incision. Ex- 
amined under the compound microscope the papilla appears as 
a minute, solid bud growing out from the terminal region of 
the radial canal, though it does not at first seem to be a direct 
outgrowth of that organ. Very soon, however, there is established 
a direct connection with the canal, and it is quite easy to dis- 
tinguish the circulation of the gastric fluid in the little bud, which 
becomes definitely vesicular, as shown in Figure 2. The growth 
of the organ, after its vesicular stage is established, is quite rapid 
and there can soon be distinguished the thickening of the terminal 
portion to form the lithocysts. Coincident with this stage of de- 
velopment there is discernible the development of the new hood 
and lappets, accessory organs, and as will be shown in connec- 
tion with a study of the histology of these organs, the correspond- 
ing development of the so-called olfactory and ocellar pits. 

In connection with the present series the following experiments 
were made with a view to demonstrate that, not only in form but 
in function, the new rhopalia were perfect organs. From one 
of the specimens just described in which three rhopalia had been 
originally excised the other five were excised on June 5th, or seven 
days after the original experiment. If the three regenerated 
organs had not yet attained to functional utility the effect of re- 
moving the others would, of course, result in the typical paralysis, 
as in the first experiment already described. As was anticipated, 
the careful removal of all the rhopalia except the three regenerated 
ones did not in the least interrupt the normal rhythm or activity 
of the creature, save to act as a stimulus to quicken it, as already 
cited in connection with a previous series. This experiment was 
repeated upon several others of this as well as subsequent series, 
and always with the same results, except in a single case which 
may as well be cited in this connection, though coming under later 

Regeneration in Rhizostoma Piihno. 


In this case the original operation had removed six rhopalia, 
leaving but two. Soon after the appearance of the new rhopalia, 
but before they had begun to approach complete development, or 
before there was any indication of the presence of lithocysts or 
pigment, the two original organs were carefully removed, and in 
this case with what might likewise have been anticipated, namely, 
the complete inhibition of the normal rhythm and the consequent 
paralysis of the organism. This inhibition continued during the 

Fig. 2. 
Section of rhopalium in early stage of regeneration, ect, ectoderm; ent, ento- 
derm ; h, hood ; mgl, mesogloea ; r. c, radial canal ; s. e., sensory epithelium. 

following two days. With the continued development of the 
new rhopalia activity was recovered, though, owing to the Inter- 
position just at this juncture of an unhealthy condition of the 
medusa, it failed to entirely recover the usual vigor or tone which 
the others had shown. 

These experiments, abundantly corroborated by subsequent ones, 
leave no shadow of doubt, it seems to me, as to the capacity of 

84 Charles W. Hargitt. 

these organisms to regenerate in the last detail one of the most 
highly specialized organs known among Coelenterata. This will 
be shown more fully in connection with the later account of the 
histology of the regenerated organs. 

Other series of experiments, continued to June 20, while varied 
In some aspects of detail, were of substantially the same character 
and with results quite similar to the preceding. 

In several of the experiments care was taken to so modify the 
form and extent of the excised portions as to secure evidence as 
to the Influence of contiguous tissues or parts upon the regenerat- 
ing organs. In Figure i is shown, for example, several aspects of 
the mode of excising the rhopalia. For the most part the excision 
was In the form of a triangular cut from the margin Inward toward 
the radial canal, as shown in the figure. The dotted line a^ will 
show also in the same connection the occasional extension of the 
cut to Include twice the usual mass. In Figure i, h, will be seen 
another form of operation. In this case the portion cut out was 

Fig. 3. 
Twin rhopalia regenerated in place of the single original one. 

rectangular Instead of triangular, as In the former. The mass 
excised in the operation also varied as before. At c, In the same 
diagram, may be seen another form of excision In which the 
cut was circular instead of angular, as in the former cases. It Is 
Interesting to note that, so far as I was able to determine, the 
form of the excision had no perceptible effect upon the form or 
rate of regeneration. In the case of the rectangular or circular 
excisions the new organ appeared in Its typical place at the median 
position of the upper portion of the notch. In the case of the 
large or small portions excised In the triangular cuts not the 
slightest difference could be distinguished. With the exceptions 
of some two or three cases to be considered, there was not the 

Regeneration in Rhizostoma Puhno. 85 

slightest evidence of any deviation from the exact position occu- 
pied by the original organ. 

The apparent exceptions referred to are as follows : First, that 
in at least two cases twin rhopalia were developed instead of the 
single original one which had been excised. This is well shown in 
Figure 3. Second, that in one case two rhopalia were regenerated 
instead of the one originally excised, but unlike the preceding, 
they appeared at different points — one in the usual position at 
the upper angle of the notch, the other at the lower, or marginal 
portion of the notch, as shown in Figure 4. 

The mere fact of the occurrence of double rhopalia during re- 
generation instead of single ones is not of itself particularly re- 
markable, for the occurrence of such features is not an unusual 
one in a state of nature, both ephyrae and adult medusae being 
occasionally found with such double organs. Some further in- 
quiry should, however, be directed to the peculiar position in 
which the organ noted in Figure 4, at a, occurs, namely, at one 
side of the notch and near the margin instead of the usual posi- 
tion. On the assumption that these organs are of sensory func- 
tion and correlated with marginal nerve centers it might be 


Fig. 4. 
Two regenerated rhopalia; a, near the margin. 

thought that in regeneration they would be likely to occur in close 
relation with such centers, and that the case under consideration 
might be thus explained. The fact is very clear, however, that such 
is not the case with the vast majority of the experiments where ap- 
parently the relation of nerve centers had nothing whatever to do 
with their position in regeneration. And when furthermore we 
reflect that these are not nervous organs in any true sense, either 
in their origin or development, though possibly correlated with 

86 Charles W . Hargitt. 

some sensory function, it must be more or less evident that such 
an explanation of the single case cited would hardly hold. 

Nor would it perhaps be more satisfactory to appeal to what has 
been designated as polarity in explaining either series. The occur- 
rence of the organs in conjunction with the radial canals and their 
apparent differentiation from terminal portions of these structures 
would seem to afford a much more probable explanation of their 
regeneration at these apparently predetermined positions. And 
may we not find in this view a simple explanation of the occur- 
rence of the anomalous case referred to in Figure 4, a, for we 
find near the margins a more or less complex network of anas- 
tomosing canals, the presence of one of which may have been 
the inciting cause of the development of a sensory body at this 
particular point. 

It is interesting to note in this connection that no appearance 
of heteromorphism occurred during the entire series of experi- 
ments. This feature I have referred to in a previous paper, '97. 
in connection with similar work on Hydromedusae. On the as- 
sumption that these organs are metamorphosed tentacles we migh 
naturally look for heteromorphic phenomena similar to that re- 
corded among the Crustacea, in which occasionally instead of an 
eye an antenna develops. Nothing of the sort, however, occurred. 
There seems in every organ and tissue a remarkably inflexible 
physiological constancy. This is the more remarkable when con- 
trasted with the highly flexible character of the polyp phase of the 
group among which are found the widest range and variety of 

The fact is not overlooked that Rhizostoma is devoid of ten- 
tacles, which might be assumed as sufficient reason why hetero- 
morphism of this sort was not manifested. The fact remains, 
however, that Its polyp has the typical tentacular equipment, and 
that In Its metomorphism they are resorbed and possibly take 
the usual course, some of them contributing toward the formation 
of rhopalia. It might be an Interesting problem to determine 
in detail just the extent of this supposed metomorphism of the 
polypal tentacles Into rhopalia. May it not be possible that the 
supposed metamorphism Is In reality a resorption and that only, 

Regeneration in Rhizostoma Puhno. 87 

and that the rhopalia are essentially independent dev-elopments 
such as are found during the process of regeneration? I merely 
raise the suggestion as it has been forced upon my attention in 
course of these experiments. It seems worth farther investigation. 
In this connection may be briefly described a phenomenon which 
only came under critical observation late in the course of the 
experiments, and which for lack of material it was impossible to 
follow out to conclusive results. Among the last of the series 
two large specimens were operated upon as follows : In the 
first all but one of the rhopalia were excised, while in the second 
all but two were removed. In both cases there was distinctly 
noticeable an aberrant, rotary sort of swimming movement, the 
animal revolving in an irregular circle, instead of directly for- 
ward or upward as is usual. Examination showed that this in- 
clination of the body in swimming was constantly in the direction 
of the remaining rhopalia, which would seem to suggest that 
perhaps they functioned something after the nature of equilibrium 
organs. I do not recall that this feature has been referred to by 
the investigators previously cited, and very much regret that it 
was not practicable for me to carry out such additional experi- 
ments as would have afforded more definite conclusions. It must 
suffice to merely mention the matter, hoping that at some time 
someone may be able to secure definite conclusions by extended 
experiments not only upon this medusa but perhaps on others 
as well. 


In connection with observations upon several specimens which 
had become degenerate or perhaps pathologic, resulting from un- 
favorable conditions of some of the aquaria, or perhaps in some 
cases due to the depleting effects of the experiments, as in the 
case of the first experiment cited in this paper, occasion was taken 
to examine somewhat in detail the observations and experiments 
of Uexkiill and to compare cases coming under my own observa- 
tions during the course of the experiments. 

In one specimen which had shown evident decline of vigor and 
upon which there appeared certain exumbrellar blotches or cor- 

88 Charles W. Hargitt. 

roslon patches, similar to those mentioned in the earher portion 
of this paper and comparable in general aspects to cases men- 
tioned by Uexkiill, it was found that after all the rhopalia had 
been removed the specimen yet exhibited certain convulsive con- 
tractions which at times simulated an irregular rhythm. I there- 
fore undertook to repeat several of this observer's experiments 
as to the effects of certain chemical stimuli, specially that of com- 
mon salt, NaCl. Small crystals of this salt were carefully placed 
on definite parts of the sub-umbrellar musculature, and I was able 
thereby to confirm in the main his results. There was a very evi- 
dent white coloration of the adjacent tissues, and this was followed 
by a more or less definite, though somewhat irregular, rhythmic 
contraction of the umbrella which continued for perhaps five 
minutes. The experiment was repeated several times and upon 
different specimens and with usually similar results, though dif- 
fering as to vigor or continuity. 

Uexkiill had concluded that the recovery of a similar rhythm 
In specimens upon which he had experimented by excising the 
rhopalia was due, not to any direct restoration of nervous or 
other normal equilibrium, but to certain pathologic conditions 
which had Intruded themselves, and among which he was spe- 
cially Impressed by these corrosion abscesses or disease patches, 
to which reference has been made. Doubting whether an agent 
of this sort, affecting particularly the exumbrella, could have any 
very definite Importance as a center of stimulus, It occurred to me 
to vary the experiment by applying the salt to the exumbrellar 
region Instead of the musculature of the sub-umbrella, and though 
variously repeated the results were uniformly negative In char- 
acter, no conclusive responses of any sort being obtained. Nor 
was there observed any of the whitening effects which were so 
evident In the previous experiments. We may conclude, it seems 
to me, that the effects produced by the salt in arousing a simulated 
rhythm of contraction was due to the direct action of the sub- 
stance on the musculature Itself, and not to any general effect 
produced upon the coordinating centers of the medusa. These 
stimulating effects of sodium chloride upon muscular tissue are too 
well known to call for any special mention In this connection. 

Regeneration in Rhizostoma Piilmo. 89 

It would seem, therefore, that in the light of these facts one 
may well question the validity of Uexkiill's conclusions, or rather 
inferences. The mere presence of whitish blotches on an organism 
would hardly justify, without the most conclusive demonstration, 
the inference that the presence of similar effects produced by some 
reagent proved them identical or even analogous. That there 
may have been certain pathologic conditions operating upon these 
medusae of which the whitish blotches were in some respects ex- 
pressions may have some measure of probability. But that these 
blotches were in themselves the inciting stimuli giving rise to the 
simulated rhythm must be regarded as doubtful, if not indeed, 
highly improbable. Such a conclusion could hardly have been 
suggested had it been observed that the same whitish blotches are 
not unusual on specimens which have been for some time in 
aquaria. Moreover, their presence on such specimens has not in 
the least, so far as my own observations have gone, served to 
Introduce any variation of the normal rhythm, a condition which 
might not be unusual on the assumption of these disease patches 
becoming sources of abnormal stimuli, and thereby Introducing 
erratic or conflicting factors into the physiological processes of 
the organism. It Is well that attention should be directed to 
disturbing conditions of this character in order that undue weight 
be not given to a single factor in determining so Important a prob- 
lem. On the other hand It may be quite as Important that In 
discrediting one conclusion there Is not substituted another of 
even less value. 

One might be tempted in this connection to go somewhat out 
of the way to consider Uexkiill's conclusions as to the purely 
mechanical function of the rhopalia In relation to the rhythmic 
action of the umbrella of medusae. If they might be supposed to 
act after the fashion of the clapper of a bell, using his figure of 
comparison, In the case of such medusae as Rhizostoma what ex- 
planation shall we have for the Identical rhythm exhibited by 
many other medusae entirely devoid of rhopalia or any equivalent 
organ? Many other objections will immediately arise when one 
reflects upon the very different histological conditions of structure 
found in these organs in various medusae, but to take up any one 

90 Charles W . Hargitt. 

of these and other phases of the problem would lead too far 
afield, and we must satisfy ourselves for the time by the reflection 
that while such speculations are interesting as well as ingenious 
they are far from demonstrations. 


A brief study of the histology of the regenerated organs shows 
the various stages of the process and establishes beyond doubt a 
true histogeny, though it has not been possible to demonstrate the 
details of mitosis in the proliferating cells. This may be due in 
part to lack of just those refinements of technique necessary to 
bring out these features. Some of the tissues were fixed by means 
of Flemming's solution, some by corrosive-acetic acid, and still 
others in lo per cent, formol in water. I have not been able 
to distinguish that there was any appreciable advantage in the 
one over the others, the formalin seeming to afford equally good 
fixation and preservation. Heidenhain's iron haematoxylin and 
an aqueous solution of haematein both afforded fairly good differ- 
entiation, though they failed as to the nervous tissues, a result 
which was not unexpected. 

In Figure 2 is shown a longitudinal section of a regenerated 
rhopalium at a comparatively early stage, when first distinguish- 
able as a minute papilla. In an earlier part of the paper I have 
referred to its early appearance as having the character of a solid 
bud from the upper angle of the notch made in the process of 
excising the organ. From an examination of this figure, which 
is among the earliest stages I have been able to satisfactorily sec- 
tion, it would seem that in its origin it probably follows the usual 
process of the regeneration or development of such organs in the 
coelenterates, namely, that of budding, involving both ectoderm 
and entoderm. As shown in the figure, there is here a typical 
outgrowth from the distal end of the radial canal and, as also 
mentioned in another connection, it was easy to demonstrate at 
about this stage of development in the living medusa an active 
circulation in the bud. The cells of the ectoderm at this stage 
are of approximately uniform size over the entire organ, and 
the same is also the case with the cells of the entoderm. There 

Regeneration in Rhizostoma Pulmo. 91 

seems also to be present the middle lamella, though less sharply 
defined than at a somewhat later period. There appears to be a 
rapid proliferation of the cells of the entoderm near the terminal 
portion where they form a mass as shown in the figure, though, as 
mentioned above, it was not possible to distinguish evidence of 

It is interesting to note at even this early stage the incipient 
phases in the regeneration of the sensory areas (Fig. 2, s. e.) just 
above and below the rhopalium. The regeneration of the hood 
is also shown at h. 

\ — eel. 


s.e. ent. 

Fig. 5. 

Section of regenerating rhopalium. rh. c, rhopalial canal ; other letters 
as in Fig. 2. 

In Figure 5 is shown a section taken in the same plane as the 
former, but at a somewhat later stage of development. The 
rhopalium has apparently attained nearly full size, but lacking as 
yet any development of otoliths, though in the network shown at 
rh. c. there is apparently evidence of a differentiation preparatory 
thereto. There is also shown here the thinning out of the ecto- 
derm of the distal portion of the organ as seen at ect. 

The sensory areas and epithelium above and below are here 
seen to have acquired almost their typical form and character as 


Charles JV. Hargitt. 

shown at s. e. The hood is also shown at h, not having appar- 
ently kept pace with the growth of the other organs. Here as 
before the direct connection of the radial canal with the rho- 
palium is quite broad and characteristic. The middle lamella, or 
mesogloea, is shown at mgl. above and below, in the latter ele- 
ments of a loose network being traceable, with the embedded cells, 
which can also be found indefinitely scattered throughout the jelly. 
In Figure 6 we have a section through an almost mature rho- 
palium, taken in the same plane as the others, only the organ 

n.f s.e. 


Fig. 6. 
Section of regenerated rhopalium, approaching maturity. It, lithocysts ; nf, 
nerve fibers ; other letters as in Figs. 2 and 5. 

itself with the terminus of the hood being shown. The ectoderm 
has become practically uniform over the entire distal portion of 
the organ, but as it approaches the base of the area of the litho- 
cysts, shown at It. and generally throughout the entire distal part, 
it becomes columnar. At 5. e. it forms a definitely arched por- 
tion, the sensory epithelmm, beneath which at n. f. is the so-called 
nerve fiber area of the nerve center of this region. While it Is 
quite possible to distinguish a more or less fibrous character as 

Regeneration in Rhizostoma Pulmo. 93 

shown In the figure, It has not been possible to trace these fibers 
into any cellular plexus, or ganglion, such as has been claimed to 
exist here. Since however my observations In the present instance 
have been almost entirely restricted to phases of regeneration, it 
will not be pertinent to discuss the question farther. 

As will be seen there Is still a continuous connection between 
the cavity of the distal portion of the organ and the radial canal. 
This connection Hesse, '95, has shown in figures of normal or- 
gans in maturity, but In the present examinations I have found it 
when fully regenerated to become entirely solid throughout the 
llthocyst region, the radial canal ending abruptly at its basal end, 
which is shown almost closed In the figure under consideration. 

In the rhopalial cavity, rh. c, which at this time Is nearly spher- 
ical, there is present a radiating network of delicate fibers, poorly 
shown in the figure, which seem to diverge from a point on the 
lower surface and extend entirely across the cavity apparently 
attaching to the opposite wall. I should consider these fibers of 
the same nature as those shown In Figure 5 near the terminus of 
the line rh. c. Though It has not been possible to critically trace 
the details of the process it seems entirely probable that the ento- 
dermic epithelium of this region becomes gradually differentiated 
Into fibers which form the Intricate network within which the 
llthocysts are later deposited. Within this network may be found 
during the various stages of development the gradual metamor- 
phosis of this entodermic cell mass, the nuclei of the cells often 
remaining as permanent elements of the organ. Some of the 
more prominent of these are shown In the figure, and phases of 
the metamorphosis may be detected near the narrow slit-like 
canal, just beyond the terminus of the radial canal. 

Within the network may also be traced the deposition of the 
pigment characteristic of the organ. 

Concerning the histology of the regenerated oral and gastric 
organs it has not seemed essential to make special Inquiry, since 
in what has already been shown In connection with the more 
highly differentiated tissues of the marginal organs It would seem 
that no serious doubt can remain as to normal histogenic pro- 

94 Charles W . Hargitt. 

cesses probably occurring throughout every regenerating organ in 
this medusa. 

It was pointed out in connection with the description of certain 
experiments that both in form and in function we have among 
the Scyphomedusae a regenerative capacity extending to the most 
highly specialized organs. In the subsequent account of the his- 
tology of the regenerated organs it has been shown that the pro- 
cess is a perfectly normal and characteristic one, conforming in 
apparently every detail to the course of development of the em- 
bryonic history of the several organs. 


EiMERj Th. — tJber Kiinstliche Theilbarkeit von Aurelia aurita und Cyanea 
capillata in Physiologische Individuen. Wurzburg, 1874. 
" Die Medusen Physiologisch u. Morphologisch auf ihr Nervensystem. 
" Tubingen, 1878. 

" Organic Evolution. English Translation. Cunningham, 1890. 
Forbes, Edw. — British Naked Eyed Medusae. London, 1848. 
Hargitt, C. W. — Recent Studies in Regeneration. Biol. Bui. Vol. I, 1897. 

" Experimental Studies upon Hydromedusae. Biol. Bui., 1899. 
Hargitt, G. T. — Notes on Regeneration in Gonionemus. Biol. Bui., 1902. 
Hesse, R. — Uber das Nervensystem u. d. Sinnesorgane v. Rhizostoma Cuvier. 

Zeits. f. wiss. Zool., 1895. 
Morgan, T. H. — Regeneration. Macmillan, 1901. 
Romanes, G. J. — Jelly-fish, Star-fish and Sea Urchins. 1885. 

Uexkull, J. VON. — Die Schwimmbewegung von Rhizostoma pulmo. Mitt. Zool. 
Sta. Neapel., XIV, 1900. 

Syracuse University, 

The Zoologicai Laboratory, 

January 20, 1904. 





With 53 Figures. 


The following observations and experiments on Leptoplana 
constitute a part of a series of investigations of form-regulation 
undertaken at the Zoological Station at Naples in 1902-3 during 
occupation of a table granted by the Smithsonian Institution. A 
part of the work on Cerianthus has already appeared (Child '03a, 
'03b, '04a.) 

The work was undertaken primarily for the purpose of examin- 
ing the relations between form-regulation and the mechanical 
tensions resulting from the creeping and other movements in some 
of the polyclad turbellaria. For this purpose is was desirable 
that the direction of movement should be relatively definite in at 
least some one of the forms studied. Many of the polyclads 
show little definiteness of direction in their movements under 
ordinary conditions. This is notably the case in the species of 
Stylochus in which, as might be expected, corresponding form- 
changes are almost completely absent. Leptoplana proved to be 
the only form readily obtainable in large numbers at Naples 
which fulfilled these conditions. The creeping movements of this 
form are relatively definite in direction. 

If form in the Turbellaria is in any way dependent upon the 
mechanical strains to which the tissues are subjected, it is to 
be expected that in forms with tough, resistant tissues, the changes 
in shape resulting from changes in mechanical conditions will 

96 C. M. Child. 

be less clearly marked than in those with relatively plastic 
tissues. As is well known, the tissues of many of the poly- 
clads are extremely tough and resistant. Stylochus, Thysanozoon, 
and many other forms might be mentioned as examples of this 
condition. As regards this feature also Leptoplana proved to be a 
favorable form since its tissues are relatively soft and plastic, 
though' much firmer than those of Planaria. With this form 
as a basis it was possible to make some very interesting com- 
parative observations upon various other species, which will be 
discussed later. 

Nearly all the specimens used were collected about the Castel 
dell' Ovo and were presumably Leptoplana tremellans. Since It 
is Impossible according to Lang ('84, p. 482) to distinguish with 
certainty the species of Leptoplana except by examination of the 
copulatory organs In serial sections, my material may have In- 
cluded other species — L. alcinoi and L. pallida. Nearly all the 
specimens used, however, resembled closely the type of L. tremel- 
laris represented In Lang's Figure i, Tafel IIL In no case did 
individual specimens exhibit characteristic differences in the regu- 
lative processes that could be regarded as specific, so the question 
as to the species Is In any event of minor importance for the pres- 
ent purpose. 

Specimens and pieces were kept Isolated or several together 
according to the experiment. In Stender dishes of various sizes, 
covered to exclude dust. The water was changed twice a week 
or oftener, but the animals proved extremely hardy and capable 
of living in small dishes even during summer for much longer 
periods without change of water. 

An attempt was made early In the course of my work to de- 
termine whether the activity of the animals was affected by light. 
So far as I could determine, specimens kept In darkness were 
slightly more active, but the difference was not sufficiently great 
to exert any marked Influence on regulation. Later most of the 
specimens were kept In darkness except when under examination. 
All were kept without food. 

Extensive series of measurements were necessary In the study 
of the form changes, and Leptoplana, like most of the turbellaria, 

Studies on Regulation. IV. 97 

Is not a favorable form for exact measurement. It was necessary 
to repeat all measurements several times In order to be certain 
that they were approximately correct. In all cases the attempt 
was made to secure the measurements while the animal was In 
the fully extended condition. In order to accomplish this It was 
often necessary to stimulate the specimens to movement and then 
to measure them while moving. A small millimeter scale which 
could be immersed In the water was used for whole animals and 
the longer pieces, while the smaller pieces were measured under a 
low power of the microscope with the aid of an ocular micrometer. 
In all cases the measurements were reduced to millimeters. 

The measurements which were commonly made are as follows : 
length of whole body, distance from anterior end of head to 
middle of group of eyes, distance from anterior end of head to 
anterior end of pharynx, length of pharynx, width of head in 
region of the eyes — In whole animals this is usually the widest 
part of the body, but In short pieces undergoing regulative changes 
of form, the widest part Is anterior to the eyes; in such cases meas- 
urement both of the widest part and the eye-region was made — and 
finally the width of the body at the posterior end of the pharynx. 
In regenerating pieces with new tissue the dimensions of the new 
tissue and the position of regenerating pharynx and genital ducts, 
If present, were also carefully determined by measurement. Since 
a cut surface undergoes marked contraction In Leptoplana and 
the new tissue arising from It consequently occupies only a part 
of the area of the original cut surface it was necessary In many 
cases to determine with great care the width of the body just 
anterior to the cut, the width of the new tissue at its origin, the 
difference of these measurements representing the degree of con- 
traction of the cut surface. Measurements of pieces without ce- 
phalic ganglia are not strictly comparable with those of pieces 
In which the ganglia are present, since the former rarely extend 
fully. It will not be necessary in most cases to give all these meas- 
urements in detail since the figures, which are drawn from them 
in almost every case, will show the changes with sufficient clear- 

98 C. M. Child. 

In all cases where the form of the pieces rendered it necessary 
figures were drawn in my notes on the basis of the measurements 
and the living specimen. By this means a record was kept not 
only of the principal dimensions, but also of any special features, 
e. g. the angle between the axis of the new and old tissues, the 
curved contours, etc. The actual form-relations and contours are, 
I think, shown by the figures as exactly as is possible in a case 
where alteration of indiv^idual form is so great. Except where 
otherwise stated the figures are about seven times the natural 
size. The various internal organs are represented so far as 
necessary in a conventional manner. 


As a preliminary to the descriptive part of the paper a brief 
discussion of certain phases of the problem in hand is desirable 
in order to clear the ground. 

In no case was the attempt made to feed the pieces employed 
for experiment. In consequence of the absence of food a marked 
decrease in size occurred during the course of the experiments. 
There is no doubt, however, that the results of feeding would be 
similar to those obtained by Morgan with Planaria (Morgan, 
'oo), for specimens were occasionally found among the worms 
collected which had regenerated after a loss of a part of the 
body. In these specimens the amount of new tissue formed was 
much greater than in the pieces kept without food, and there is 
no reason to believe that in pieces where regeneration is possible 
growth to the full size may not occur, provided enough material 
is at hand. As Morgan ('98 ) has pointed out, the material used 
in the formation of new tissue in starving pieces is obtained from 
the substance of the piece itself or from its reserve supplies, and 
the bulk of the old tissue is reduced to a greater or less extent 
by the formation of the new tissue. When the pieces are fed 
the amount of new tissue formed is more or less increased and 
the old tissues not only do not decrease in size, but may grow 

The fact that formation of new tissue may occur, not only 
once but repeatedly, in pieces which have been for weeks without 

Studies on Regulation. IF. 99 

food indicates that the stimulus which brings about the regen- 
eration is sufficiently powerful to deprive the old portions of ma- 
terial. There is little doubt that this difference indicates a dif- 
ference in metabolic activity between the old and the growing 
regions. If this conclusion be correct it follows that in the pres- 
ence of nutrition the new parts will grow more rapidly than the 
old. Moreover, we find that in the absence of food growth of 
the new tissue may cease long before the amount of tissue removed 
has been replaced. We must conclude therefore either that the 
stimulus to regeneration decreases as regeneration proceeds, or 
that the old portions give up material less rapidly as the process 
continues (Child, '03d). That there is an actual difference in 
quality between the new and old tissue is clearly shown by ob- 
servations which I have made repeatedly, viz., that in regenerat- 
ing pieces kept without food until death occurs from starvation, 
infection or other causes, the old parts usually disintegrate before 
the new. In many cases I have seen the old tissue disintegrate al- 
most completely in pieces of Leptoplana, while the new tissue re- 
mained alive and apparently healthy for a considerable time after- 
ward. Moreover, the nev/ parts in regenerating specimens show a 
greater degree of muscular and other functional activity than do 
the old parts. 

For reasons which I hope to state in full at some future time, I 
believe that this difference is at least in part dependent upon 
functional conditions, viz., that it concerns the use or ac- 
tivity of the parts. In the earlier stages of regeneration 
other factors are very probably concerned in greater or less 
degree. The presence of a cut surface places the cells adjoining 
it under conditions widely different from those existing before 
the cut was made. The equilibrium in physical conditions is de- 
stroyed by the removal of the part and the absence of pressure 
from other parts on one side may itself be sufficient to bring about 
a migration or growth of tissue outward from the cut surface. It 
is extremely difficult in such cases to determine how much of the 
" new tissue " is the result of migration and how much of actual 
proliferation. But if the attempt is made by the animal to use 
the ' new tissue " thus formed in the manner characteristic of 

loo C. M. Child. 

the part which it represents new conditions of pressure and ten- 
sion arise as well as powerful nervous stimuli which probably 
affect growth and differentiation directly or indirectly. 

When for instance the posterior part of the body of one of 
these worms is removed the animal continues to move about and 
" attempts " to carry out the same movements as when the pos- 
terior end was present, but in the absence of the parts the move- 
ments fail more or less completely of success. In fact observation 
of these cases leads me to believe that in the absence of the part 
the attempts to attain the usual result are often more powerful 
than when it is present. For example, a specimen of Leptoplana 
with the tail removed makes violent attempts to hold to the 
substratum by the cut posterior end of the body as well as by 
other parts; a specimen with the lateral lobes of the head removed 
makes violent but unsuccessful attempts to swim; and finally, to 
take another case somewhat removed from the present considera- 
tions, a fish with the tail or part of it removed uses the remaining 
stump much more vigorously than would be the case if the whole 
were present. 

It may appear at first glance that these statements involve 
unwarranted assumptions regarding the psychological activities of 
forms as low in the scale as the Turbellaria, but I believe such a 
conclusion is not justified. Conscious recognition of the successful 
or unsuccessful character of the movement is by no means neces- 
sary, but on the other hand it is diflicult to understand how these 
creeping worms could advance in a regular, definite manner if 
the movement over the substratum or the movement of the parts 
of the body upon each other did not afford certain characteristic 
stimuli. The movements of these forms as well as those of higher 
animals are coordinated, and for coordination some stimulus re- 
resulting from the movement seems to be necessary. Removal of 
a part, e. g., the posterior end by which the animal has been ac- 
customed to attach itself, must bring about a change in the rela- 
tion of the various stimuli. The animals behave in such cases 
as if they were moving over surfaces to which their bodies do 
not adhere readily. They appear to make violent efforts to use 
the parts which are'missing. These changes in behavior are dis- 

Studies on Regulation. IV. lOi 

tinct from any irritation due to a wound, for they continue after 
the wound has closed and new tissue has appeared. There is 
no doubt, I think, that a modification of the motor stimuU occurs 
in the absence of a part important to locomotion. 

The outgrowth of new tissue from the cut surface is probably, 
in its earlier stages, the result of the alteration in local conditions 
consequent upon the removal of a part. But the position of the 
new tissue, i. e., its connection with a particular part of the old 
body determines the conditions to which it is subjected in connec- 
tion with the functional activities of the old differentiated parts. 
As differentiation in the new tissue proceeds, motor activity ap- 
pears and soon the movements of the new part are coordinated 
more or less completely with those of adjoining old parts. Thus 
the conditions to which the new part is subjected become similar 
to those which were present in the part removed. These condi- 
tions or some of them are undoubtedly formative factors in many 
cases. In Stenostoma (Child, '02, '03) the development of the 
tail depends in large degree upon their presence. 

Now when the new part first shows characteristic coordinated 
motor activity it is much smaller than the part removed, yet func- 
tionally it supplies the place of the other, though at first very 
imperfectly. But the smaller the size and the more imperfect 
the formation of the new part, the greater the activity, i. e., the 
"attempt" of the animal to use it. Thus the new part is visibly 
more active than the old and if we admit that the conditions con- 
nected with this activity are "formative factors" it is easy to see 
why in a starving piece the new part continues for a longer or 
shorter time to increase in size at the expense of the old tissue. 

As the new part increases in size and its coordinations become 
more perfect the degree of motor activity decreases, approaching 
that of the old parts. The extent of regeneration in starving 
pieces is probably determined by the relative functional activity 
of the new and old parts. As long as the more intense metabol- 
ism of the new part enables it to deprive the old part of material, 
so long will it continue to increase in size. It is also probable 
that the old tissue gives up material less and less readily as the 
encroachments continue. The final result depends on the condi- 
tions of the individual case. 

io2 C. M. Child. 

It is possible that the effect of the functional conditions may 
be in many cases largely mechanical, i. e., that in consequence 
of the use or attempt at use of a growing part, e. g., a 
regenerating tail, it is subjected to certain mechanical condi- 
tions of tension and pressure and that these mechanical condi- 
tions themselves constitute in reality the chief "formative factor," 
acting either mechanically or as physiological stimuli to growth. 
In many cases, however, there is no doubt that other internal 
stimuli bring about growth, but even in such cases mechanical 
conditions must usually play a certain part in the final arrange- 
ment of the material produced. In short there must usually and 
perhaps always be a mechanical factor of more or less import- 
ance in regulative morphogenesis. I think it probable that in 
the lower animals this mechanical factor is relatively simple but 
of great importance, while with increasing complexity it becomes 
more complex and more difficult of analysis, though perhaps not 
less important. 

The alteration in general outline and proportion of pieces, 
especially of the old portions, called by Morgan morphallaxis, 
which occurs during regulation in such forms as Planaria (Mor- 
gan, 'oo, 'oi), Stenostoma (Child, '02, '03) and Leptoplana, I 
believe to be primarily due to mechanical factors connected with 
locomotion and acting very probably both in a simple mechanical 
manner and as stimuli to growth, though there is some reason 
to believe (Child, '02) that the direct mechanical effect is pre- 
dominant in many cases. We cannot conclude, however, that all 
phenomena which have been designated as morphallaxis are due 
to similar conditions. The changes in form of pieces of the 
medusa Gonionemus for instance (Morgan, '99) cannot be due 
to the factors which cause the change of form in Stenostoma and 
Planaria, but are very probably due to physical conditions in the 
tissues whose equilibrium is destroyed by a removal of a part, 
and so may be comparable to the inrolling which occurs in pieces 
of Cerianthus (Child, '04a). In dealing with problems of so 
great complexity generalizations are safe only so far as the actual 
facts go. Nothing is gained by referring these diverse phenom- 
ena to an inherent capacity in pieces for returning to the original 
form. Such an explanation leaves us exactly where we started. 

Studies on Regulation. IV. 


If these views are correct it follows that these form-changes, at 
least in the old parts, and often in the new as well, must occur 
to a greater extent when regeneration is not quantitatively com- 
plete, or must be more evident, since the growth of the 
parts may mask it to a greater or less extent. This is actually 


Fig. 1. 
the case, as Morgan's experiments have shown (Morgan, '00). 
Since my primary object in investigating the regulative processes 
in Leptoplana was the examination of the alterations in propor- 
tion, the most favorable* conditions for this purpose, viz., absence 
of food, were desirable. 

I04 C. M. Child. 

On the other hand the study of regeneration in the stricter 
sense is not at all impossible under those conditions. The failure 
of the new portion to attain full size is a minor matter. Indeed 
the presence of food is a complicating factor in the study of reg- 
ulation of lower forms, since it renders less possible the distinc- 
tion between ordinary processes of growth and the regulative 


In Leptoplana, as in Stenostoma (Child, '02, '03a, '03b), 
there is a close relation between form-regulation and movement. 
A description of the characteristic methods of movement may 
properly, therefore, precede the account of experiments. 

Locomotion in Leptoplana tremellaris is accomplished in two 
ways, by swimming and by creeping. Lang ('84, pp. 634-636) 
has described the movements of the polyclads and among them 
those of Leptoplana. I desire, however, to consider these move- 
ments with special regard to their mechanical effect upon the 
tissues and for this purpose Lang's description does not suffice. 
Figure I shows the outline and proportion of a specimen in fully 
extended condition as when creeping. 

Swimming is accomplished by an undulating movement, dorso- 
ventrally directed, proceeding posteriorly from the anterior end 
of the lateral regions of the head and anterior portions of the 
body, the median portions remaining meanwhile almost motion- 
less. This method of swimming is called by Lang the flying 
movement. In various other polyclads it appears in much more 
extreme form than in Leptoplana and in some involves not only 
the anterior regions but the whole lateral region of the body as 
in Thysanozoon. 

It is interesting to note that in all cases where this undulating 
movement extends over only a part of the lateral region of the 
body, the region involved is the broadest portion of the body. 
In Leptoplana it is not sharply marked off from other regions 
posterior to it, as is the case in some forms, e. g., Stylochoplana 
agiUs (Lang, '84, Fig. 2, Tafel II, also pp. 457 and 636). In 
correspondence with the absence of sharp demarcation of the 
undulating region in Leptoplana we find that the undulating 

Studies on Regulation. IV. 105 

movements do not cease abruptly as they pass posteriorly but 
gradually decrease in amplitude until no longer visible. During 
extreme activity they may extend much further posteriorly than 
under ordinary conditions and frequently slight undulations of 
the margins appear along the sides and pass posteriorly even 
when the animal is creeping. During swimming the anterior 
region of the body is considerably broader than in Figure i. 

It can scarcely be doubted that these movements play a part 
in shaping the regions in which they occur. A comparison be- 
tween frequency, amplitude, and force of the undulating move- 
ments and the degree of lateral development in the regions in 
which they occur is most striking. According to the usual point 
of view this correlation between structure and function is merely 
one of the many remarkable cases of adaptation, but in my opin- 
ion it is, at least in part, the direct result of function in the indi- 
vidual. Some experimental evidence bearing on this point will 
be offered elsewhere. 

As regards the manner in which the movement may affect the 
tissues it is not difficult to see that the movement of these parts 
to and fro through the water must subject them to tension in the 
in the lateral direction. This must affect in greater or less degree 
the distribution and arrangement of the plastic tissues composing 
the parts. A very simple physical experiment serves to illustrate 
this point. A cylindrical or square stick of sealing-wax moved 
to and fro in one plane in water sufficiently warm to soften it 
will undergo flattening in a plane at right angles to the direction 
of movement. The change in form is more strikingly shown if 
a rigid axis is present; a mass of wax molded in cylindrical form 
about a stiff wire will become in a few minutes a thin, flat plate 
decreasing in thickness towards the edges and with a rounded 
outline. The mechanical conditions resulting from the move- 
ment of the wax through the water are not widely different from 
those which the undulating margins of Leptoplana produce. If 
the wire axis of the wax be considered as the longitudinal axis the 
effect of movement through the water is lateral extension. In 
Leptoplana the undulating movement is confined chiefly to the 
lateral regions in the anterior third of the body and it follows that 
the conditions described are limited chiefly to these parts. 

io6 C. M. Child. 

There can be little doubt, in my opinion, that these mechanical 
conditions constitute a factor in the formation of the broad lateral 
regions in Leptoplana and more especially in other forms in which 
the undulating movements of these parts occur. In other words 
the form is in some degree the result, not the cause, of the char- 
acteristic method of activity. The experimental data to be de- 
scribed support this view. 

In addition to its power of swimming, Leptoplana is able to 
creep over surfaces rapidly and in a definite direction. Both 
muscular and ciliary activity are concerned in the movements, 
but one or the other may predominate according to conditions. 

When the animal is moving quietly, as for instance after a 
slight stimulation, the cilia afford the chief motive power, although 
the slight muscular movements of the margins of the body are 
almost constant, portions being lifted from the substratum, 
brought forward, and again attached. This muscular play of 
the margins is especially marked in the anterior regions but extends 
in some degree along the whole side of the body. 

After a strong stimulus the movements take on a different char- 
acter, becoming chiefly muscular. The portions of the body In 
which the undulating movements occur during swimming furnish 
under these conditions the chief motive power. Parts of the 
margin are lifted slightly, extended in the antero-lateral direction, 
and attached to the substratum : contraction of the muscles fol- 
lows and the body is drawn forward. These movements occur 
in rapid alternation on the two sides of the body and the similar- 
ity between this mode of progression and the use of legs can- 
not escape the observer. The animals appear almost as if walk- 
ing forward. 

At all times during creeping movements the body adheres 
closely to the substratum as may be demonstrated by sudden at- 
tempts to dislodge it. The chief regions of attachment are the 
lateral margins and the posterior end. Frequently during creep- 
ing small portions of the body margin which adhere more closely 
than other parts are stretched posteriorly to a considerable degree 
before they are torn away from the substratum. As In many 
other Turbellaria the posterior end is an important organ of 

Studies on Regulation. IV. loy 

attachment although in Leptoplana it is not so exclusively em- 
ployed for this function as in many other forms. 

Leptoplana differs from many other species of polyclads in the 
definite direction of its movements. In some forms, e. g. Stylo- 
chus, the direction of movement is very indefinite, movements in 
other directions being almost as frequent as anteriorly directed 
movements. In Leptoplana, however, the deviation from the 
longitudinal direction is slight. 

As a general rule the more posterior portions of the margin 
and the posterior end itself are used more frequently as organs 
of attachment than the more anterior regions. 

In consequent of the adhesion to the substratum by the margins 
and posterior end, the body of Leptoplana is subjected to mechan- 
ical tension in the longitudinal direction, often visibly in a con- 
siderable degree, during creeping. As in the case of Stenostoma 
(Child, '02, '03), this longitudinal tension constitutes a factor 
in determining the general form and outline of the body. The 
fact that the margins as well as the posterior end are employed 
as organs of attachment accounts for certain characteristic fea- 
tures in connection with the form. 

From the facts above cited regarding movement we must con- 
clude that the posterior portions of the body are subjected more 
frequently than the anterior parts to longitudinal tension In con- 
sequence of their poslton and more frequent use for attachment, 
and moreover, that the tension is greater than that In the anterior 
regions since all ciliary impulses and muscular contractions aiding 
in forward movement anterior to the point of attachment com- 
bine to produce it. This statement is correct in a simple case 
but frequently various points along the margin may become at- 
tached simultaneously and the tension Is distributed among them. 
The continual muscular play of the margins, the rapid transitions 
which a given region undergoes from attachment to reattachment 
are of course accompanied by great variation in mechanical con- 
ditions. The Important point is that the tissues are subjected to 
longitudinal tension and the posterior regions more than the an- 

io8 C. M. Child. 

This case differs from that of Stenostoma in which the posterior 
end alone is the chief organ of attachment. Reference to my paper 
on Stenostoma (Child, '02) will show clearly how the character- 
istic differences of external form betewen Stenostoma and Lepto- 
plana may be correlated with the differences in the mechanical 
conditions to which the tissues are subjected. 

My experiments also indicate that the use of the anterior por- 
tions of the lateral margins in drawing the body forward consti- 
tutes a factor in their development. In consequence of these char- 
acteristic, frequently repeated movements these parts are sub- 
jected to characteristic physical conditions, which, like the longi- 
tudinal tension, must exert some influence upon the arrangement 
of the cells and tissues. 

Anyone who observes the creeping movements of different poly- 
clads cannot fail to note the close correlation between the gen- 
eral outline of the body and the character of the movement. In 
general the forms which advance in a definite direction are more 
slender than those like Stylochus whose movements are very in- 
definite in direction. In the last mentioned form lateral move- 
ment occurs almost as often as longitudinal, a part of the body- 
margin being advanced and the other portions drawn up to it by 
contraction. The breadth of the body is almost as great as the 
length in Stylochus. I am forced to the belief that the forms of 
the various species are determined in greater or less degree by the 
conditions of tension, resulting from swimming and creeping 
movements, to which the tissues are subjected. The experiments 
to be described afford strong support to this view. 


This section includes merely a brief preliminary statement 
concerning the power of regeneration in Leptoplana. The phe- 
nomena will be treated more at length in other connections. 

Complete anterior regeneration never occurs in Leptoplana 
when the cephalic ganglia are removed. Removal of all portions 
of the head anterior to the ganglia and even including the anterior 
part of the ganglia Is followed by rapid and complete regenera- 

Studies on Regulation. IV. 109 

When the cut is made at any level posterior to the ganglia 
neither the ganglia themselves nor the head are regenerated (Cf. 
Lillie, '01). 

Posterior regeneration is qualitatively complete at all levels 
posterior to the ganglia whether the ganglia are present or absent 
in the piece, but pieces cut anterior to the ganglia never regenerate 
the ganglia nor the posterior parts. 

Lateral regeneration is qualitatively complete when the ganglia 
are present, but when they are absent neither they nor the lateral 
part of the head removed are regenerated though lateral regen- 
eration of other parts may be more or less complete in the ab- 
sence of the ganglia. Removal of the right or left half of the 
ganglia is followed by complete regeneration from the remaining 

In general the amount of tissue regenerated in pieces kept with- 
out food is much less than that removed, though all the organs 
may be present. The amount of posterior regeneration varies 
inversely as the distance of the cut surface from the anterior end. 
The size of the piece does not affect the quality of regeneration 
and affects the amount only slightly, except on approach to the 
minimal size, when a marked decrease in the amount of regenera- 
tion occurs. The minimal size of pieces capable of qualitatively 
complete posterior regeneration was not determined with exact- 
ness, but transverse pieces less than one tenth the length of the 
body are still capable of qualitatively complete posterior regenera- 
tion and pieces even smaller than this, but containing the cephalic 
ganglia, regenerate completely in all directions. 


Considering first one of the simplest cases, viz., posterior re- 
generation from a transverse cut surface we find that in Lepto- 
plana, as in Planaria and other Turbellaria, the new tissue which 
makes its appearance on the cut surface assumes a rounded outline 
and grows or extends posteriorly in the direction of the longi- 
tudinal axis, becoming more slender and tapering as regeneration 
proceeds. Figures 2 — 4, drawn from careful measurements, will 
serve as an illustration of the course of regeneration in such cases. 

I lO 

C. M. Child. 

The cut surface in this piece was a short distance posterior to thc^ 
cephalic ganglia. Figure 2 represents the piece five days after 
section, Figure 3 ten days after section, and Figure 4 twenty-seven 
days after section. The new tissue is bilaterally symmetrical at 
all times and growth appears to occur most rapidly along the me- 
dian plane. The gradual decrease in size of the whole is due of 
course to the absence of food. The course of posterior regenera- 
tion in Planaria, as described by Morgan and others, is similar, 

Fig-. 2. 

Fig. 3. 

Fig-. 4. 

F^g. 5. Fig. 6. .Fig. /, 

though the amount of new tissue formed is relatively less than in 
Leptoplana; similar results have also been obtained by others 
with various forms. 

If the cut surface from which regeneration occurs be oblique 
instead of transverse the course of regeneration differs in some 
respects from that just described. Figures 5-7 illustrate the his- 
tory of such a piece, begun on the same day as the pre- 

Studies on Regulation. IV. 1 1 1 

ceding and examined at the same Intervals. Figure 5 shows the 
piece five days after section, Figure 6 ten days after section, and 
Figure 7 twenty-seven days after section. In Figure 5 the new 
tissue Is symmetrical with respect to the contracted cut surface 
but not with the median plane of the animal. As growth proceeds 
however, a gradual change In direction of the axis of the new 
tissue occurs (Figure 6), until finally this corresponds with the 
median plane and approximate bilateral symmetry of the whole 
results, though the new tissue Is still unsymmetrical In form since 
the surface from which It arose Is oblique. 

This change in the direction of regeneration Is also familiar 
to students of regeneration, having been described by Morgan 
and others for Planaria and other forms. It seems to bear the 
stamp of a true regulative process for it brings the parts Into the 
position which they must occupy In order to produce a bilaterally 
symmetrical whole. 

During observations on Planaria in which the change Is well- 
marked, the possibility suggested itself that it was primarily due, 
not to some Internal factor operating In such manner as to pro- 
duce the typical form of the species or an approximation to it, but 
rather to the locomotion of the animal in the direction of the lon- 
gitudinal axis. It appeared probable that since the new parts 
were used for attachment and thus subjected to tension in the di- 
rection of the longitudinal axis they were gradually drawn out In 
this direction and so a symmetrical whole was produced. This 
view was supported by the fact that the change seemed to begin 
when the new part became functional. When the new tissue first 
appears In these forms It Is apparently little used for attachment 
or at least without complete success. Within a few days, how- 
ever, the specimens can be seen to adhere closely to the substratum 
by means of It, and it is at this time that the apparent change In 
the direction of growth first becomes conspicuous. 

The question as to the effect of altering the direction of loco- 
motion In pieces at once presented itself to me and fortunately I 
found In Leptoplana a favorable form for experiments of this 
kind. Short pieces from the body of Leptoplana containing the 
cephalic gaoglia or a considerable portion of them move In circles 

I 12 

C. M. Child. 

when one side of the body is cut away, since the axis usually be- 
comes bent and there is nothing to counterbalance the effect of the 
cilia and muscular movements of the opposite side. The results 
of experiments with pieces of this kind demonstrated in a most 
satisfactory manner the correctness of my belief. Numerous ex- 
periments were performed, the results in all cases being unequivo- 
cal. In the following sections some of these experiments are de- 



A very satisfactory method of obtaining pieces which move in 
curves is that of separating the anterior end by a cut a short dis- 
tance posterior to the cephalic ganglia and splitting this piece lon- 
gitudinally in half at or near the median line. This method of 
preparation is illustrated by Figures 8 and 9. Figure 8 shows the 

Fig. 9. 

direction in which the cuts are made and Figure 9 the piece after 
contraction of the cut surfaces has taken place. It is evident from 
the latter figure that the contraction is an important factor in 
bringing about circular locomotion. The longitudinal axis of the 
piece becomes bent toward the cut side and movement in a straight 
line is impossible. The curve of locomotion approximates more 
or less closely the curve of the axis but does not necessarily coin- 
cide with it since the irregular form of the piece often alters the 
direction. The tissue giving rise to the new posterior region soon 
begins to show the effect of the direction of movement and the tail 
forms at an angle with the old parts. The description of the fol- 
lowing series will serve to illustrate the course of regeneration. 

Studies on Regulation. IV. 


I. August 31, 1903. A specimen of average size was prepared 
as shown In Figure 10. The greater part of the body was re- 
moved by a transverse cut about 2 mm. posterior to the cephalic 
gangha and the anterior piece thus obtained was split longitudin- 
ally. In this case the longitudinal cut appeared to be coincident 
with the median plane, but the differences In behavior of the two 
pieces after section Indicated that the left cephalic ganglion was 
injured to a greater extent than the right. After section the cut 

Fig. 10. Fig, 11. 

Fig. 13. 

Fig. 16. 

Fie- 15. 

Fig. 17. 

Fig. 18. 

surfaces soon contracted thus bending each piece Into a curved 
form resembling Figure 9 and In both pieces locomotion diverged 
constantly toward the cut side, the pieces thus moving in circles, 
as Indicated by the arrows accompanying the figures. 

September 3: 3 days after section: 

The pieces have assumed the forms shown In Figures 1 1 and 
12. New tissue has appeared almost uniformly over the whole 
extent of the cut surface. The left plecp (Figure 12) is consid- 
erably more contracted than the right piece (Fig. 11) and moves 

114 CM. Child. 

somewhat more slowly. Numerous experiments to be discussed 
later have shown that with increasing injury to the cephalic gan- 
glia the rapidity of movement and the degree of extension de- 
crease, hence it is probable that the ganglion in the left piece has 
suffered greater injury than that in the right. 

September 6 : 6 days after section : 

At this stage the regenerating posterior end has become distinct 
and the first traces of the new pharynx are visible (Figs. 13 and 
14). The posterior outgrowth is directed toward the cut side 
and Its axis coincides with the curve of locomotion. The new 
tissue is now functional to some extent, the tail being employed 
by both pieces for attachment. 

September 20: 20 days after section: 

The two pieces are shown in Figures 15 and 16. The axis of 
the new body is distinctly curved in both and the pharynx shows 
in each case some degree of curvature. In each the part of the 
ganglia removed is regenerating and In connection with It are a 
few eye spots. The small new ganglion was distinctly visible in 
the living specimens from the dorsal side. Both pieces continue 
to move in circles, though with a somewhat larger radius than be- 
fore, the change being due to the development and use of the new 
tissue along the side of the head, which now evidently aids In loco- 
motion and thus counterbalances in some degree the effect pro- 
duced by the old parts. Portions of the margin of both the new 
tissue and the old can be extended antero-laterally and attached 
to the substratum, and tension Is exerted upon the other parts by 
muscular contraction of these regions. But the power of the new 
portions Is still much less than that of the old parts. Similarly, 
in consequence of the curvature of the old parts, itself due to the 
contraction following section, the effect of the cilia on these parts 
is such as to cause movement in a curved line which Is not yet 
counteracted by the cilia on the new parts. 

The larger right piece was accidentally Injured at this time and 
its fupther history could not be observed. 
October 12 : 42 days after section: 

Figure 17 represents the left piece at this stage. Considerable 
reduction in size has occurred, but the amount of regenerated tis- 
sue is relatively much greater than in Figure 16. The direction 

Studies on Regulation. IV. 115 

of movement diverges less from a straight line than before, and 
correspondingly the angle between the longitudinal axis of the head 
and new body is decreasing. The new cephalic ganglion is nearly 
as large as the old, but the eyes are still less numerous in the new 
tissue than in the old. The regeneration of the lateral regions 
of the head has proceeded so far that anterior to the eyes the new 
portion is nearly as broad as the old. 

The change in form of the regenerating lateral margin of the 
head is the most conspicuous feature of this stage (compare Figs. 
16 and 17) . It has now acquired almost its typical form. More- 
over, the curvature of the longitudinal boundary between the new 
and old portions, /. e. the longitudinal cut surface, is decreasing. 
Observation of the movements of this piece at this stage showed 
that the functional activity of the regenerated margin of the head 
was very great. It was much used in locomotion, portions being 
extended anteriorly or antero-laterally, attached, and then con- 
tracted, thus drawing the body forward. Swimming movements 
were also often made, though short pieces of this kind do not suc- 
ceed in swimming to any extent, being apparently unable to main- 
tain their equilibrium in the absence of posterior parts of normal 
size. There can be little doubt that the functional activity of 
this region has brought about the change in form. Characteris- 
tic movements have produced a characteristic arrangement of the 
tissues. Moreover, the frequent extension of the margin anteri- 
orly followed by attachment and contraction has undoubtedly 
aided in forcing the anterior part of the old tissue toward the left 
and thus straightening the outline of the cut surface. 

October 24: 54 days after section: 

As indicated in Figure 1 8 the changes described above con- 
tinue. The piece moves still more nearly in a straight line than 
twelve days ago and the form is correspondingly altered. The 
present stage exhibits one interesting effect of the change in direc- 
tion of the tension upon the tissues. During locomotion the 
margin at a", including both new and old tissue, is thrown Into 
small wrinkles or folds, while the right side of the body Is very 
evidently stretched. The folds are clearly the result of the 
altered direction of tension In the adjoining parts. The body 
grew out in a direction differing considerably from that in which 

ii6 C. M. Child. 

it extends at present and with the change in position the tension 
on the tissues at this point has decreased until now it has become 
pressure and these parts are "too long" for the position they must 
occupy under the altered conditions, A comparison of Figure i8, 
the form of the piece during locomotion, and Figure 19, the form 
during rest, when the parts are not subjected to longitudinal ten- 
sion, renders it still more evident that the tension due to move- 
ment is the cause, not the effect of the change in form. When the 
piece is at rest the angle between the original axis and the axis of 
the new body is always greater than during locomotion and the 
folds at X disappear. In other words the change in direction of 
the new body does not precede but follows, and does not even keep 
pace with the change in direction of locomotion. These facts 
leave no room for doubt that the tension due to locomotion is the 
efficient factor. 

In a previous section the fact was noted that locomotion in Lep- 
toplana is chiefly ciliary when the animal moves quietly, but that 
when strongly stimulated the movements are to a large extent 
muscular. The same is true of these pieces. When stimulated only 
slightly they progress at a uniform rate, largely by means of the 
cilia, but under stronger stimulation the margins of the head are 
used in the manner described and the body is drawn forward by 
strong muscular contractions usually alternating on the two sides. 
A marked difference in direction between the two kinds of loco- 
motion was observed in this piece and indeed in many other similar 
pieces. The direction of locomotion by muscular contraction was 
sometimes after strong stimulation in a curve to the left while that 
of ciliary locomotion was always toward the right. During loco- 
motion the muscular activity of the right side of the head — the 
new tissue — appears to be greater than that of the old tissue on 
the left. In ordinary locomotion the muscular play of the mar- 
gins of this part is much more conspicuous than on the left. Ap- 
parently the new parts are in a more active condition functionally 
than the old, and doubtless under strong stimulation are capable 
of more work. When the piece turns to the left after strong 
stimulation the difference in muscular activity between the two 
sides Is much more marked, that of the right side being clearly 
much greater. 

Studies on Regulation. IF. 


November 8: 69 days after section: 

At this time the locomotion of the piece diverges only occasion- 
ally and then slightly from a straight line, except sometimes after 
strong stimulation when the piece turns to the left. The form is 
still more nearly symmetrical (Fig. 20). The folds which were 
visible at x in Figure 18 have disappeared in consequence of rear- 
rangement or resorption of the superfluous parts (atrophy from 
disuse?), though when the piece turns to the left after strong 
stimulation folds appear temporarily In this region. 

The longitudinal boundary between new and old tissue is now 
almost a straight line, i. e. the contraction of the cut surface which 
occurred after removal of the right side is now scarcely percept- 

Fig. 22. 

Fig. 21 

Fig 23 


Fig. 24 

Fig. 25. 

Fig. 26 


Fig. 27. 

ible. This change Is probably, like the contraction Itself, due 
primarily to an alteration in mechanical condition, i. e. the changes 
in mutual pressure and tension. 

II. August 31, 1902. The body was removed by a transverse 
cut about I mm. posterior to the cephalic ganglia and the anterior 
piece tlius obtained was split longitudinally (Fig. 21 ) . The lon- 
gitudinal cut passed a little to the left of the median plane and so 
through the left cephalic ganglion. In consequence of the con- 
traction of the worm during the operation the course of the cut 

ii8 C. M. Child. 

was curved as shown in the figure. Only the piece on the right 
of the cut will be considered here. After section the cut surfaces 
contracted and the anterior end bent over so far that the outline 
of the anterior region became almost symmetrical (see the outline 
of the old tissue in Fig. 22). 

September 3 : 3 days after section : 

Figure 22 shows the piece as it appears at this stage. The 
contraction has brought the two cut surfaces, originally at right 
angles into almost the same plane. New tissue has begun to ap- 
pear but there is no marked difference in amount in different re- 
gions. That portion of the left cephalic ganglion which remained 
in the piece protrudes slightly from the cut surface and is indicated 
in the figure by deep shading. The piece moves in rather small 
circles as indicated by the arrow. 

September 6: 6 days after section: 

As indicated in Figure 23 the new tissue, probably the ectoderm, 
has united with the protruding portion of the left cephalic gan- 
glion and is thus prevented from extending at this point. An- 
terior and posterior to this region growth has occurred and in a 
curious manner. It appears as if two posterior ends were forming, 
one from the lateral region of the head, the other from the poster- 
ior cut surface, both of them corresponding in direction to the ten- 
sion resulting from locomotion. Both adhere to the surface to 
some extent, but the posterior one somewhat more firmly. Several 
factors combine to produce this peculiar condition : the piece has 
contracted In such a manner that the cut surface from which the 
left side of the head would normally regenerate faces somewhat 
posteriorly; the union between the protruding nervous tissue and 
the new tissue divides the growing region into two parts; and 
finally the tissue representing the left side of the head is just be- 
coming functional so that its margin reacts to the contact of the 
substratum but does not yet extend anteriorly and contract strongly 
and aid in locomotion, being instead stretched postero-laterally 
since It adheres to the substratum until the forward movement 
loosens It. This condition shows very clearly how effective me- 
chanical tension may be as a ''formative factor." 

Studies on Regulation. IV. 119 

September 10: 10 days after section: 

Unfortunately the condition described above did not continue, 
for the ganglionic mass became separated from the right ganglion 
and remained united with the new tissue on its dorsal surface near 
the left margin (Fig. 24). In consequence of this change the 
regions of new tissue before separated are now continuous and the 
outline of the margin is rapidly undergoing alteration. More- 
over, the new lateral tissue is now further developed functionally 
and its margin reaches forward, attaches itself and contracts in 
the characteristic manner, this assisting in the locomotion which 
consequently becomes slightly less curved in direction. Corres- 
ponding with this increase in characteristic functional activity is 
the convex outline of the lateral margin of this region. As long 
as it was being subjected to postero-lateral tension this portion of 
the margin was in part slightly concave like the sides of a growing 
tail or body ( Fig. 23 ) . 

The mass of ganglionic substance affords a landmark which 
enables us to determine that the terminal portions of the body are 
formed first, a conclusion agreeing with that of various authors 
in regard to soft parts at least in other forms. 

September 20: 20 days after section: 

Figure 25 represents the condition at this stage. The direction 
of movement is still far from a straight line, though the curvature 
is decreasing. The curved body and pharynx require no special 
comment. The new lateral region of the head now functions 
very activ-ely and a comparison of Figures 23, 24 and 25 shows 
that the curve of contraction of the original cut surface is becom- 
ing less marked, i. e. the old tissue is being pressed back toward 
the right at the anterior end by the active new tissue. Regenera- 
tion of the left cephalic ganglion is taking place. 

October 12: 42 days after section: 

The new lateral region of the head is now so active that it 
counterbalances the old part to a considerable extent and the di- 
rection of movement is less curved. Figure 26 shows the speci- 
men at this stage. The change in form and the growth anteriorly 
of the lateral region is marked (compare Figures 25 and 26). 
Small folds at x during locomotion indicating the pressure exerted 

120 C. M. Child. 

by the functional activity of the new part. The angle between 
the body and the original longitudinal axis is decreasing, i. e., the 
body is swinging into typical position. 

October 24: 54 days after section: 

At this stage the direction of locomotion approaches still more 
closely a straight line, and the form is correspondingly changed 
(Fig. 27). The small folds at x due to the pressure of the new 
lateral region against the old parts at the anterior end are still 
visible, and similar folds appear at xx in consequence of the change 
In position of the body. The left margin of the head shows great 
activity In the region where the lateral outgrowth Is greatest, and 
frequently performs swimming or "flying" movements of some 
amplitude. The old portions, on the other hand, are less active. 
The regenerated cephalic ganglion is nearly as large as the other 
and eyes are present In connection with it. 

Loss of the piece a few days later prevented completion of the 

Fig. 29. 

Fig. 28. 

Fig. 30. ^^i-21- 

III. August 31, 1902. A specimen was prepared in the manner 
described for Series I and II, the longitudinal cut being made 
as nearly as possible in the median plane. Probably, however, 
it was actually a little to the right of the median plane, since, as 
In Series II, a part of what seemed to be the right cephalic gan- 
glion protruded from the cut surface of the left piece, the part 

Apparently the left cephalic ganglion was more or less injured 
by the operation, for during the first two weeks the piece showed 

Studies on Regulation. IV. 121 

little motor activity. As regeneration proceeded, however, it be- 
gan to revolve in circles scarcely greater in diameter than its own 
length, appearing almost as if revolving on a pivot. 

The extreme curvature of the direction of locomotion renders 
the piece of interest and certain points require consideration. 
During the first period after section when locomotion was slight 
regeneration was almost uniform over the whole cut surface, 
i. e., there was no marked extension of the portion representing 
the posterior region. Figure 28 represents the piece ten days 
after section. A comparison of this figure with Figure 24 the 
corresponding stage of Series II In which active locomotion has 
occurred during the ten days suggests the possibility that the 
tension or other conditions connected with locomotion may not 
only determine the direction of outgrowth of new tissue but may 
also affect the amount of regeneration, a point which will be dis- 
cussed more fully elsewhere. 

Within the next few days the piece began to move in the manner 
described above and twenty days after section appeared as repre- 
sented in Figure 29. The axis of the regenerated portion forms 
an angle of more than ninety degrees with the original longitud- 
inal axis. The protruding mass of nerve tissue has united with 
the new tissue and delayed growth at that point. The rapid 
change of form which has occurred in the new tissue during ten 
days indicates the marked effect which use of the parts exerts 
upon regeneration. 

Twenty-two days later, forty-two days after section, the form 
was much the same. Figure 30 represents the piece during or- 
dinary locomotion. Frequently the piece assumed the form shown 
in Figure 3 1 in which the tip of the tail was overlapped by the 
lateral margin of the head. Figure 32 represents the form as- 
sumed when the piece attached Itself by the tail and contracted, 
drawing Itself backward; In this condition folds appear at x, 
indicating that pressure Instead of tension occurs In that region. 

About ten days later the piece died without further changes. 
This case differs from the preceding in that no marked change in 
the direction of the body-axis occurred during the whole history, 
although the piece lived as long as many others in which the 

122 C. M. Child. 

change occurred. The continued circular locomotion is of course 
directly responsible for the absence of change in form and this 
in turn may be due to the delayed regeneration of the right ce- 
phalic ganglion and consequent imperfect coordination of the 
new lateral margin. As a matter of fact the right lateral margin 
of the head appeared much less functionally activ^e than in the 
other cases described. Regeneration of the ganglion was delayed 
by the presence of the old ganglionic tissue which did not lose its 
connection with the left ganglion until about four weeks after sec- 
tion. And finally this long-continued attachment of the injured 
ganglionic tissue to the left ganglion is doubtless to be ascribed 
to the fact that scarcely any locomotion occurred during the first 
two weeks after section, so that the new tissues with which the 
ganglionic tissue was united were not subjected to tension which 
would aid in removing this tissue from the region where its pres- 
ence interfered with regeneration. I have no doubt that had the 
piece lived sufficiently long before exhaustion occurred, the right 
lateral margin of the head would have acquired its characteristic 
activity and so would have counterbalanced the motor effect of 
the old tissue, thus bringing about the change in direction of the 
body-axis which occurred in Series I and II. 

In all of the cases described thus far the transverse cut surface, 
originally posterior, becomes oblique in consequence of the con- 
traction and in most cases the outgrowth of new tissue forming 
the body occurs in a direction nearly perpendicular to this surface, 
though there is considerable variation in different cases. Thus 
in Figures 15 and 16 of Series I the angle between the axis of 
the new body and the cut surface is somewhat more than 90°, 
while in Figures 23, 24 and 25 of Series II it is approximately 
90°, and in Figures 28, 29 and 30 of Series III It is again more 
than 90°. These cases therefore are open to the objection that 
the direction of growth may have been determined in some degree 
by the direction of the cut surface rather than by the tension due 
to movement, for it is a well known fact that in many cases regen- 
eration takes place chiefly at right angles to the cut surface. Al- 
though I did not consider this objection valid I prepared other 
series in which the posterior cut surfaces of the pieces were 

Studies on Regulation. IV. 


strongly oblique (Fig. 33) in order to obtain experimental evi- 
dence on the question. A few cases from one of these series are 
described. These cases show that the direction of regeneration 

Fig. 33. 

Fig. 39. 

Fig. 42 

Fig. 34. 

Fig. 35. 

Fig. 40. 

Fig. 41. 

Fig. 45. 

Fig. 43, Fig. 44. 

is not determined, except of course in the early stages, by the sur- 
face or surfaces from which it occurs. 

124 C". M. Child. 

IV. September 3, 1902. Four large specimens were cut in the 
manner represented in Figue 33. By this method the relation 
between the plane of the cut surface and the direction of locomo- 
tion is different in the right and left pieces. 

Two pieces of each set are described, the others showing no 
additional features of importance. 

1. A piece from the left side (see Fig. 33) : Figure 34, seven 
days after section; Figure 35, twenty days after section; Figure 
36, thirty days after section. In this case the cut surfaces re- 
mained nearly in their original relations and the piece was not 
greatly bent. Consequently the curvature of the axis and of the 
direction of locomotion was not as great in many cases. The 
angle at which the body appears corresponds with the direction 
of locomotion, but the axis of this region is far from perpendicular 
to the posterior cut surface. In later stages this specimen became 

2. A piece from the left side: Figure 37, seven days after sec- 
tion; Figure 38, twenty days after section; Figure 39, thirty 
days after section. This piece became so bent during contraction 
that the posterior cut surface faced somewhat toward the right 
instead of to the left as originally (compare Figs. 33 and 37) and 
the direction of locomotion was correspondingly curved. The 
longitudinal cut was a little to the left of the median plane, thus 
injuring the left cephalic ganglion to some extent. The piece 
was consequently less active In locomotion and the right margin 
did not acquire full functional activity as soon as In many other 
cases. There was therefore no marked change in the direction 
of locomotion and the curvature of the axis persisted to a great 
extent up to the time of death. In this case also It Is evident that 
the outgrowth forming the posterior region Is not perpendicular 
to the posterior cut surface. 

3. A piece from the right side of the head (see Figure 33) : 
Figure 40, seven days after section; Figure 41, twenty days after 
section; Figure 42, thirty days after section. In this case the 
contraction of the piece, though not great, brought the two cut 
surfaces almost Into line. The curvature of the regenerating 
body is clearly shown in Figure 41. The outgrowth is more 

Studies on Regulation. IV. 125 

nearly perpendicular to the posterior cut surface in this case than 
in the two preceding cases, but this is to be expected from the 
position of the latter. In this piece the regeneration of the left 
cephalic ganglion was not delayed, the left margin of the head 
became functionally active within a month after section (Fig. 
42) and reduction of the curvature began and was completed 
before death. 

4. A piece from the right side of the head: Figure 43, seven 
days after section ; Figure 44, twenty days after section ; Figure 
45, thirty days after section. The longitudinal cut in this case 
injured the right ganglion to some extent. The piece became 
greatly bent and simply revolved within a space little greater than 
its own size. The posterior cut surface was brought into line 
with the longitudinal surface. The posterior region grew out 
toward the anterior tip of the head, but not at right angles to the 
plane of the cut surface. Within the month the left cephalic 
ganglion was partially regenerated and the left margin of the 
head attained some degree of functional activity thus reducing 
the curvature of locomotion and the axis of the body began to 
straighten (Fig. 45). The piece did not, however, attain any- 
thing like symmetrical form before death. 

A comparison of these four cases renders it sufficiently evident 
that the angle between the plane of the posterior cut surface and 
the regenerating body may vary greatly, while, on the other hand, 
the relation between the direction of the outgrowth and the direc- 
tion of locomotion is evident. 

In the cases described the relation between the functional ac- 
tivity of the regenerating margin of the head and the direction of 
locomotion has been pointed out. Attention has also been di- 
rected to an apparent relation between the development of this 
functional activity and the regeneration of the cephalic ganglion 
of that side. The relation of the nervous system to regeneration 
will be discussed elsewhere, but the fact may be noted here that 
the presence or absence of the one ganglion appears to affect the 
regeneration of the lateral margin of the head, but not that of 
the posterior region. The latter may be formed in the usual man- 
ner when only one ganglion is present, though it is usually not as 

126 CM. Child. 

long under these conditions as when both are present. I think 
there is little doubt that the difference is connected with the func- 
tional activity of the parts. Each region develops its character- 
istic form only as it is used in the cjiaracteristic manner. In all 
cases marked growth anteriorly and laterally of the new margin 
of the head has been observed as soon as the animal begins to use 
it in the manner characteristic of these regions, while the case 
represented in Figure 23 shows that so long as this region is not 
used in the ordinary manner it may develop posteriorly. The 
posterior region of the body can perform its functions to some 
extent in the absence of the cephalic ganglia as will be shown else- 
where, and moreover, in the cases under consideration the regen- 
erating posterior region is undoubtedly innervated from the part 
of the nervous system present. It therefore performs its usual 
functions, though perhaps less perfectly, before the other ganglion 
regenerates, and its development proceeds in the typical manner. 
To put It briefly, the margin of the head develops a characteristic 
form because used in a characteristic manner, and the body de- 
velops a different form because it is used differently. The preced- 
ing experiments are sufficient to show that mechanical tension Is 
an important factor In morphogenesis in these animals. No one 
would admit more readily than myself, however, that many other 
factors may be concerned here and that in other cases the factors 
may be wholly different. 

Attention has been called In several cases to the apparently 
greater functional activity of the regenerated parts as compared 
with the old In later stages. This difference indicates, I believe, 
a real physiological difference. The old portion decreases in size 
In consequence of loss of material while the new parts increase 
In absolute size in the earlier stages and in relative size in later 
stages. It is not at all Improbable that the functional condition 
of the reduced old part differs widely from that of the new part. 
The former, reduced to a fraction of Its former size, Is certainly 
less plastic mechanically and probably less sensitive to stimuli. 
The new part Is to be regarded as possessing the qualities of a 
young and growing organism, the old on the other hand as ap- 
proaching exhaustion. The difference observed in functional ac- 

Studies on Regulation. IV. 127 

tivity between new and old portions agrees well with the fact 
already mentioned in the section on "Regulation, Nutrition and 
Use of Parts," that in starving pieces the old part usually dies and 
disintegrates before the new, which may live and remain appar- 
ently healthy for several days after the loss of the old part. 


Another method employed for bringing about circular loco- 
motion was that of making oblique cuts at various levels posterior 
to the cephalic ganglia. The results obtained by this method are 
in some respects less striking than those described in the preceding 
section, but since the cephalic ganglia are not injured in any way 
by this method, a possible objection to the preceding series is ren- 
dered invalid. 

The change in direction of the longitudinal axis of the regen- 
erating tissue arising from a posterior oblique cut surface has 
been mentioned (see also Figures 5-7). In this case the direc- 
tion of locomotion was not altered by the cut, and, as might be 
expected, the axis of the new tissue soon became coincident with 
that of the old. In the course of similar experiments I found, 
however, that if the cut were very oblique the contraction of the 
cut surface following the operation might bring about circular 
locomotion. Figure 46 represents the manner in which such a 
cut is made, and Figure 47 the piece after contraction. It Is 
evident that contraction produces a marked curvature In the lon- 
gitudinal axis, and therefore the piece In advancing turns con- 
stantly toward the cut side. It was found necessary to make the 
cut in the anterior region of the body in order that markedly 
circular locomotion might occur, because if the cut were made 
near the middle of the body or in the posterior half, the part 
of the body in which the axis was not affected by the cut was so 
long that the bilaterally symmetrical Impulse to movement from 
this region nearly or quite counterbalanced the effect of the bent 
portion, and the regenerating tissue grew out in the direction of 
the old axis. In all cases where the circular locomotion was well 
marked the cut was made either just anterior to the pharynx as 


C. M. Child. 

Fig. 4.6. 

Fig. 49. 

Fig. 4/. 

Fig-. 50. 

Fig. 48. 

Fig. 51. 

Fig. -52. 

Fig. 53. 

Studies on Regulation. IV. 129 

In Figure 46 or through Its anterior portion. Within certain 
hmlts the circular locomotion Is more marked as the obliquity of 
the cut Increases, the reason being clear. If, however, the cut 
be nearly longitudinal the slender strip on the longer side Is likely 
to roll up and may act as a drag, thus complicating locomotion 
and delaying or preventing typical regeneration. 

Figures 48-53 Illustrate the history of a piece cut in the manner 
indicated in Figure 46 and moving In a curve toward the right. 
Even as early as three days after section (Fig. 48) the new tissue 
is symmetrical with respect to the cut surface, evidently In con- 
sequence of the effect of locomotion. Figure 49 shows the con- 
dition six days after section. Here the curvature of the new 
tissue is becoming evident. In Figure 50 — tw^elve days after 
section — the curvature of the regenerating part Is still more con- 
spicuous. Twenty-eight days after section the tissue has acquired 
the form shown In Figure 51. Now that the regenerated tissue 
has attained a considerable length and its posterior region con- 
stitutes the chief organ of attachment some degree of straight- 
ening occurs. The manner in which this takes place is shown in 
Figures 52 and 53, both of which represent the same stage — fifty- 
eight days after section. Figure 52 shows the piece In ordinary 
locomotion. Here the margins as well as the tip of the tail, or 
frequently only the margins or certain regions of them, adhere to 
some extent, and the resulting tensions cannot cause straightening 
of the longitudinal axis since they follow approximately the same 
curve. Frequently, however, only the posterior end of the body 
adheres, or the margins of the head and the posterior end, and at 
such times the piece assumes temporarily the form of Figure 53, 
in which the axis is nearly straight. The longitudinal tension to 
which the body Is subjected straightens it, but at the same time 
bends the contracted posterior part of the old tissue to the left so 
that the outline becomes convex at the left of this region. At the 
same time small folds appear at x, indicating that In this region 
the tissues are subjected to pressure instead of tension. This posi- 
tion Is never maintained for any length of time, and as soon as 
the tension ceases the piece resumes the form of Figure 52. There 
can be no doubt, however, that if this position is taken sufficiently 

130 C. M. Child. 

often straightening will occur. This I believe is the chief factor in 
the change from the curved to the straight bilaterally symmetrical 
form. This piece died after sixty-six days without having become 
completely symmetrical. 

The history of other pieces prepared in a similar manner Is 
essentially the same, though, as has been mentioned, the degree of 
curvature of the new tissue varies within certain limits with the 
obliquity of the cut and with the level at which the cut is made. 
The reason for variation in curvature with the angle of the cut 
lies in the fact that the more oblique the cut the more the axis of 
the piece is bent and consequently the greater is the curvature 
of locomotion. 

As regards the level, the curvature of the new tissue decreases 
as the distance of the cut from the anterior end increases, because 
the region of the body in which the axis is not bent by the con- 
traction of the cut surfaces increases and counteracts the asym- 
metrical motor effect of the bent region more and more com- 
pletely. This variation of curvature with the level of the cut 
renders it evident that the cut surface itself has little influence 
upon the direction of growth except in the earliest stages, for a 
piece cut at a given angle near the anterior end, e. g. as in Figures 
46-53, will give rise to new tissue with a marked curvature, while 
in another specimen cut at the same angle but farther posteriorly 
the new tissue will show much less curvature while in still another, 
cut at the same angle in the posterior pharyngeal region, the 
new tissue will grow out in the direction of the longitudinal axis 
or will very soon acquire this direction, simply because the direc- 
tion of locomotion is curved only very slightly or not at all. 
In each case the direction of outgrowth coincides with the line 
of locomotion, 


The bearing and the significance of the experiments described 
is sufficiently clear to render extended discussion unnecessary. They 
may, I think, be regarded as demonstrating the fact that the ex- 
tended form of the body in Leptoplana is determined in large 
degree by mechanical conditions. It is difficult to describe ac- 

Studies on Regulation. IV. 131 

curately the continually changing movements of these animals, but 
I think no one who actually observes the pieces can fail to be 
convinced of their importance in determining form. It is true of 
Leptoplana as of Stenostoma (Child, '02) that it has no "normal 
form," t. e., no definite hard and fast form inherited and devel- 
oping In the Individual Independent of physical conditions. What 
these animals do possess Is a capacity for certain kinds of activity. 
These are given potentially in the chemical and physical structure 
of the protoplasm, which to my mind represents rather capacity 
for functional activity in the broadest sense than form. As my 
experiments prove, certain elements of form In the morphological 
sense develop incidentally as the result of functional activity In 
in a given environment. These elements have been commonly 
regarded as typical and determined by heredity because they are 
common within certain limits of variation to all individuals of 
a species, but when we consider that under natural conditions 
both functional activity and environment are essentially similar 
in different individuals of the species the reason for likeness In 
these form-elements becomes clear. It is only when we can alter 
the functional activity as I have done experimentally in the case 
of Leptoplana, or the environment, as I succeeded in doing for 
Stenostoma (Child, '03a) that the dependence of these elements 
of form upon these two factors becomes clearly evident. 

But these experiments concern only morphological characters 
of a certain kind. Experiments of others have already shown that 
"formative factors" are many and various, and generalizations 
from the consideration of a single group of characters are unsafe. 
It win never be possible to explain form on the basis of a single 
principle. All the complex activities of which organisms are 
capable are "formative factors" : when we can view all of these 
In their complex interrelations and know the part which each 
plays, then and only then shall we "understand" organic form. 

The relation between form and heredity has never been satis- 
factorily determined. With the advance in our knowledge the 
fact becomes more and more evident that the organism is not 
merely a complex of structural elements ready made by heredity 
for certain functional activities, but rather a complex of ac- 

132 CM. Child. 

tlvlties in consequence of which morphological structure develops 
Physical and chemical structure of protoplasm must not be con 
fused with morphological structure : the distinction between the 
two is important though often overlooked. As regards the indi- 
vidual the former represents capacities for activity, i. e., for 
transformation and transference of energy, or in short, functional 
activity in the broadest sense. Form in the morphological sense, 
is the combined result of this activity and the erxvironment, ex- 
ternal or internal. According to this view, it is functional capacity 
that is inherited rather than form : heredity is, strictly speaking, 
a physiological and not a morphological problem. 


1. Locomotion in Leptoplana is accomplished by two meth- 
ods, swimming and creeping. In swimming the lateral regions 
of the anterior part of the body perform undulating movements 
in a dorso-ventral direction. Creeping movements are both ciliary 
and muscular, the muscular movements consisting chiefly of an 
alternate extension anteriorly or antero-laterally of the margins 
of the head, adhesion to the substratum and muscular contrac- 
tion, thus drawing the body forward. In creeping the margins 
and posterior end of the body are used as organs of attachment. 

2. In consequence of the typical movements the tissues of 
the body are subjected to typical mechanical tensions and pres- 
sures which constitute formative factors. 

3. The effect of these mechanical conditions upon the tissues 
is at least in part directly mechanical, but they may also act as 
physiological stimuli to growth (formative stimuli). 

4. The effect of the mechanical conditions incident to locomo- 
tion may be demonstrated experimentally by various methods. 
The method described in this paper consists in making the pieces 
of such a form that the direction of locomotion becomes curved 
instead of straight. In these experiments the regenerating part 
grows in the direction of the principal tension, even though this 
form an angle of 90° with the typical direction of growth. 

5. The experiments lead to the conclusion that in Leptoplana 
the regions of the body develop in a characteristic form because 
they function or attempt to function in a characteristic manner. 

Studies on Regulation. IV. 133 


Child, C. M., '02. — Studies on Regulation. I. Fission and Regulation in Ste- 
nostoma. 4, Archiv. f. Entwickelungsmech., Bd. XV., H. 2 & 3, 1902. 
'03a. — Studies on Regulation. II. Experimental Control of Form-Regu- 
lation in Zooids and Pieces of Stenostoma. Archiv f. Entwick- 
elungsmech., Bd. XV., H. 4, 1903. 

Child, C. M., '03b. — Studies on Regulation. III. Regulative Destruction of Zo- 
oids and Parts of Zooids in Stenostoma. Archiv. f. Entwick- 
elungsmech., Bd. XVII., H. I, 1903. 

'03c. — Form-Regulation in Cerianthus. I. The Typical Course of Re- 
generation. Biol. Bull. Vol. v., No. 5, 1903. 

'03d.— Form-Regulation in Cerianthus. II. The Effect of Position, Size, 
and other Factors upon Regeneration. Biol. Bull., Vol. V., No. 
6, Vol. VI., No. I, 1903. 

'04a. — Form-Regulation in Cerianthus. III. The Initiation of Regenera- 
tion. Biol. Bull., Vol. VI., No. 2, 1904. 

Lang, A., '84. — Fauna und Flora des Golfes von Neapel. XI. Die Polycladen. 
Leipzig, 1884. 

LiLLiE, F. R., '01. — Notes on Regeneration and Regulation in Planarians. Amer. 
Journ. of Physiol., Vol. VI., No. 2, 1901. 

Morgan, T. H., '98. — Experimental Studies of the Regeneration of Planaria 
maculata. Archiv. f. Entwickelungsmech., Bd. VII., H. 2 & 3, 1898. 

'99. — Regeneration in the Hydromedusa Gonionemus vertens. The 
Amer. Nat., Vol. XXXIIL, 1899. 

'00. — Regeneration in Planarians. Archiv. f. Entwickelungsmech. Bd 
X., H. I, 1900. 

'01. — Regeneration. New York, 1901. 




T. H. Morgan. 

It has long been known that the pollen of some plants will 
not fertilize the ovules of the same plant. The cause of this 
impotence has not yet been detected. 

It is also known that pollen from another plant is often pre- 
potent in those cases where normal self-fertilization may occur. 
It has further been shown, especially by Darwin, that the offspring 
from self-fertilized ovules are in general not so vigorous as those 
from cross-fertilized ones. 

There are here two problems, which, even if they should prove 
to be fundamentally related, can be most profitably examined 
separately; — first, the problem of the Inability of the male ele- 
ment to fertilize the female germ-cells of the same individual; 
and, second, the effect of self-fertilization (in those cases In which 
It occurs) on the offspring. Both problems appear to be within 
the range of experimental examination. 

There are only a very few cases known amongst animals where 
conditions similar to those in plants have been found to prevail, 
although very few hermaphroditic animals appear to have been ex- 
amined in this respect. Close inbreeding, which is commonly sup- 
posed to bring about deterioration in some cases. Is perhaps not 
very dissimilar to self-fertilization. Whether in the case of in- 
breeding there is ultimately a loss of power to fertilize the egg, 
or whether the egg fails to develop after It has been fertilized, 
has not, so far as I know, been determined. 

Castle discovered in the ascidian, Ciona intestinaUs, a case ap- 
parently similar to those in plants. The eggs are generally In- 
capable of self-fertilization, yet can be readily cross-fertilized; i.e., 

136 T. H. Morgan. 

the spermatozoa of an Individual will not fertilize the eggs of 
that individual, but have the power to fertilize the eggs of any 
other individual. 

My object in undertaking a study of this problem was, in the 
first place, to determine if possible the nature of the conditions 
that prevent or interfere with self-fertilization; and in the second 
place, I was not without hope of being able to find some way 
in which self-fertilization could be artificially Induced. As will 
appear in the sequel, these two questions are not two sides of 
the same problem ; for, while it has been possible to discover the 
means of bringing about self-fertilization, it still remains to be 
definitely determined what conditions In the egg normally prevent 
the entrance of the spermatozoa of the same Individual. 

Since Castle's observations had shown that the ascidlans offer 
favorable material for a study of this sort, I first turned my 
attention to this group, using the three most available species 
found at Woods Hole, or in the vicinity; namely, Ciona intes- 
tinalis, Molgula manhattensis, and Cynthia partita {Styela sp.). 
The work was done while holding the Bryn Mawr Table at the 
Marine Biological Laboratory, from June to September, 1903. 
Owing to the scarcity of Ciona I have not been able to work out 
completely a number of Important problems connected with one 
of the two main questions that I examined. In the near future 
I shall hope to complete this side of the investigation. 


The ovary of Ciona Is a sac-shaped body of fair size lying 
loosely attached in the coil of the Intestine. It can easily be 
removed without cutting Into the testis. Its lumen contains some 
of the ripe eggs, but the majority of these are in the oviduct. 
The oviduct can readily be opened and the eggs set free without 
cutting into the vas deferens, which follows a course parallel to 
the oviduct. If the animal is kept Isolated for 24 hours the ovi- 
duct becomes greatly distended with eggs, and after another 24 
hours even more eggs may have accumulated. The eggs are 
laid normally In the early morning, at dawn, and Castle has re- 
corded that Ciona deposits its eggs and sperm with the regularity 

Self -Fertilization Induced by Artificial Means. 137 

of the rising sun. The rough handling incidental to removal and 
isolation appears to cause Ciona to retain its eggs for several 
days. The individuals to be used were isolated, as a rule, from 
24 to 48 hours, and in most cases were rinsed in fresh water before 
opening. It was not found necessary to boil the water; for check 
experiments "showed that eggs left to themselves were never fer- 
tilized by stray spermatozoa in the sea-water. Since Ciona de- 
posits its eggs only in the very early morning, the chances are 
very slight that functionally active spermatozoa would be present 
in the sea-water in the late morning and in the afternoon when 
the experiments were carried out. 

The eggs of Ciona are surrounded by a rather thick membrane. 
Standing out like broad spikes over the surface of the membrane, 
and forming a beautiful aureole around the egg, are the trans- 
parent follicle cells, each with a shining drop in its outer end. 

A'number of preliminary experiments confirmed Castle's con- 
clusion that self-fertilization is rarely possible in Ciona intes- 
tinalis. The evidence, however, on which Castle based this con- 
clusion is not altogether satisfactory, since he records many cases 
in which self-fertilization occurred. Instances are cited in which 
isolated individuals gave 90, 25, 16, 5, 4, o per cent, of self- 
fertilized eggs. Castle supposes that, in the first of these cases 
at least, the spermatozoa of one day fertilized the eggs of the 
next, but it has not been shown that the spermatozoa have this 
power if left so long in sea-water. The same individuals that 
had been used for these isolation experiments were killed (after 
being washed in 90 per cent, alcohol), and the eggs and sperm 
of each taken out and mixed together. The results gave 50, 4, 
I, Ys, o per cent, of self- fertilized eggs. The same exj^rriment 
repeated with fresh individuals gave 50, 12^, 10, 5, 2, o per cent, 
of self-fertilized eggs. From these figures it is clear that in some 
cases a considerable amount of self-fertilization occurred, unless 
there was some source of error in the experiment. In fact. Castle 
believes that in those cases where a large number of eggs were 
fertilized there was some contamination. My own results with 
Ciona have never given so large a percentage of self-fertilized 
eggs, and I am inclined to attribute this result in part to the 

138 T. H. Morgan. 

precaution that I took to isolate the individuals the day before 
they were to be used. I have rarely seen more than from i to 10 
per cent, of self-fertilized eggs segment, and in the greater num- 
ber of cases not a single egg segmented. On the other hand I 
found, as did Castle, that as a rule 100 per cent, of cross-fertilized 
eggs develop, to which statement I should add, provided the 
spermatozoa are in "good" condition. 

What is the meaning of these remarkable facts? Why do 
not the sperm fertilize the eggs produced by the same individual, 
and yet fertilize those of any other individual? A number of 
possibilities readily suggest themselves, and since the following 
pages record an attempt to test these suggestions they may be 
briefly mentioned here : 

1. That the spermatozoa are not made sufficiently active by 
secretions from the eggs of the same individual, but by those from 
the eggs of any other individual. 

2. That the spermatozoa are not "attracted" to the eggs of the 
same individual. 

3. That the egg contains, or secretes some substance that les- 
sens the activity of the spermatozoa of the same individual. 

4. That some mechanical difficulty prevents the spermatozoon 
from entering the egg of the same individual. 

5. That even if the spermatozoon enters, it can not fertilize 
the egg of the same individual, in the sense of causing the egg 
to begin to develop. 

In order to discover if the lack of power to self-fertilize the 
eggs is due to the absence of some substance around the eggs that 
excites the spermatozoa, the following experiment was carried 
out. The eggs of an individual (A) were taken from the ovi- 
duct. Similarly the eggs of another individual (B) were also 
taken out. Then the ovary of (A) and that of (B) were 
crushed separately, and a little sea-water was added. The eggs 
of (A) were then allowed to soak in the crushed ovary extract 
of (B) and those of (B) in the extract of (A). After a short 
time the sperm of (A) with a little water was added to the (A)- 
eggs, and the sperm of (B) to the (B)-eggs. If the sojourn 
of the eggs in the extract of the ovary of another individual has 

Self-Fertilization Induced by Artificial Means. 139 

the postulated effect, or if the presence of the extract of the ovary 
of another individual has the postulated effect on the sperm, fer- 
tilization ought to have occurred. The results showed, however, 
that fertilization did not take place. 

This experiment was performed four times, giving eight sets 
in all. In six of these sets not a single egg segmented. In two 
others a very few eggs segmented (6 per cent, in one, 5 per cent, 
in the other), but this sometimes occurs in self-fertilized eggs 
not treated in any special way. Moreover there may have been 
contamination in the latter case. 

Another experiment similar in some respects to the last was also 
carried out. The heart of one individual was opened and the 
blood collected. The eggs of another individual were put into 
this blood and allowed to stand. Later, sperm of the same indi- 
vidual was added in sea-water, but no fertilization occurred in 
one set and only one per cent, in the other. Check eggs were also 
kept in this experiment to make certain that no sperm had acci- 
dentally gotten into the blood. That none were present was 
shown by the fact that no fertilization took place. It is evident 
from this experiment that self-fertilization can not be brought 
about by soaking the eggs in the extract from the ovary or in the 
blood of another individual, although the somewhat high per- 
centage of self-fertilized eggs that segmented in two cases after 
treatment with the ovarian extract may have resulted from the 
influence of the extract on the spermatozoa. 

If the spermatozoa are excited to greater activity by the pres- 
ence of the eggs of another individual it seemed not improbable 
that this might be directly observed. Therefore, I placed some 
of the sperm with the eggs of another individual and more 
of the same sperm with the eggs of the same individual, and 
compared the two preparations under the microscope. The sper- 
matozoa of Ciona are not very active as a rule, nor do they 
accumulate in crowds around the eggs, as they do in many other 
animals, or at least not to any marked extent. It seemed to me 
in both cases that sometimes the spermatozoa were more active 
immediately in the vicinity of the eggs, And in the spaces between 
the follicle cells, but as they also show the same activity around 

140 T. H. Morgan. 

pieces of the tissue of the body of the same or of another In- 
dividual I have not laid much stress on this observation, or ac- 
credited the results to the presence of an exciting substance. At 
times I have thought that the spermatozoa were more active 
around the eggs of another Individual than around the eggs of the 
same Individual, but as there Is no very accurate means of determin- 
ing their relative motility, unless very marked, I should not wish, 
as yet, to give a final answer to this question. It is certain that there 
Is no such great difference In the behaviour of the spermatozoa 
In the presence of the eggs of the same and of another individual 
as to suggest that the difference In the result Is connected with this 
factor. And even if this were the case, the Influence probably 
extends for only a short distance from the surface of the egg, as 
the following experiment shows. 

The eggs were taken from the oviduct, great care being taker 
not to injure the sperm-duct. The eggs from another Individual 
were collected In the same way. An equal number of eggs from, 
each were put together and fertilized with the sperm from one 
of the individuals. In another dish another lot of the same eggc 
were mixed half and half, and these fertilized with the spern 
from the other Individual. In each of these two sets half at 
least of the eggs should be fertilized by the other sperm, but half 
should not be fertilized unless the eggs of one Individual exer. 
some Influence that causes the sperm to fertilize the eggs of the 
same individual also. It was found that only about half of the 
eggs were fertilized. This result shows that the fertilization I 
probably not due to some substance set free by the eggs that act? 
on the sperm or at least that if such a substance is set free Its 
action Is confined to the Immediate vicinity of the egg. The 
experiment does not show, however, whether the egg, or its 
membranes, may not contain some substance that prevents the 
spermatozoa from entering the eggs of the same Individual. Ever 
If such a substance Is set free from the eggs it may not have had 
time In my experiment to accumulate sufl'iclently In the surround- 
ing water to have prevented the spermatozoa from fertilizing the 
other eggs, which may be quickly entered. This view can be 
tested by letting eggs stand In a small amount of water for c 

Self -Fertilization Induced by Artificial Means. 141 

long time, then taking out some sperm from the same individual, 
first making it active by placing it in sea-water, and then putting 
it into the water in which the eggs have stood. On the hypothesis 
these sperm should soon be brought to rest, and if then the eggs 
of another individual are added, they should not be fertilized, o 
at least not in the same proportion as when the sperm is taker 
directly from the oviducts, put into sea-water, and then added t 
the eggs. 


My first experiments with ether were made in order to deter- 
mine whether when eggs are etherized it might not be possible to 
self-fertilize them. The results turned out somewhat differently 
from what I had anticipated, for although I found that it was 
possible to self- fertilize the eggs in ether-solutions, the result 
seemed to be due to the action of the ether on the sperm rather 
than on the eggs. 

The experiment was first made with Cynthia, which in most 
cases has very sluggish spermatozoa. I observed that the first 
effect of the ether was to make the sluggish sperm very active, and 
even greatly quickened the activity of already active sperm. Fur- 
thermore I found that spermatozoa that scarcely moved at all 
in sea-water became active in the ether-solutions. Finally I found 
that in ether-solutions of certain strengths the eggs of Cynthia and 
of Ciona could be self-fertilized. The eggs behave in this respect 
so capriciously that I was obliged to carry out a large number of 
experiments in order to determine the conditions that lead to the 
self-fertilization of eggs in ether-solutions. The outcome was 
only partially satisfactory, but the experiments opened up a field 
for research, in which it may be possible to obtain further results 
of interest. 

The experiments with ether were carried out as follows : At 
first I used a nearly saturated solution of ether and diluted it a 
half, or a fourth, etc. In the later experiments I used solutions 
of known strength. It was found by trial that the solutions were 
effective between 0.25 and 5 per cent. Some of the results may 
now be given in detail. 

142 T. H. Morgan. 

Experiment I. The eggs were removed from an individual that 
had been isolated 20 hours. The sperm was also taken out, and, 
together with the eggs, was put into ether-solutions, 5, 2, i, 0.7, 
0.5 per cent, in sea-water. After 5 minutes, and again after 10 
minutes, the eggs were removed to pure sea-water. The eggs 
were injured by the ether in the strongest solution, but neverthe- 
less one segmented. In all of the other solutions about 80 per 
cent, of the eggs divided; the most in the weaker solutions. 

Experiment II. In this experiment the eggs and the sperm 
were put together into ether solutions of 5, 2, i, 0.7, 0.5 per cent. 
Some of the eggs were transferred to water after 5 and 10 min- 
utes, but others were left in the solutions. In the strongest solu- 
tion the eggs w^ere killed. In the others the following results were 

Eggs segmented Eggs segmented. 

Ether 5 minutes in ether. 10 minutes in ether. 

2. percent. 20 per cent. 25 per cent. 

I. " 5 " 75 " 

0.7 " 2 " 2 

0.5 " 10 " 5 

.J't is clear that the stronger solutions gave the best results, and 
that ten minutes immersion was better than five minutes. None 
of the eggs that were left in the ether-solutions segmented. This 
does not mean that they were not fertilized, but that the ether so 
injured the eggs after a long immersion, that they failed to de- 
velop. Several check experiments were also made in this case. 
In one the eggs were not self-fertilized but were put into a 5 
per cent, ether-solution, and transferred after ten minutes to 
sea-water. They did not segment, nor did a few that were left 
behind in the ether solution. In another check series the eggs 
were not fertilized, and were left in sea-water. None segmented, 
which shows clearly that the ether in the preceding experiment 
was in some way responsible for the self-fertilization of the eggs. 
It should also be recorded that tadpoles developed from all the 
fertihzed eggs that had been in the ether-solutions. 

Experiment III. The eggs and sperm of the same individual 

Self-Fertilization Induced by Artificial Means. 143 

were put into ether-solutions of 3, 2, i, 0.7, 0.5 per cent., and 
were removed to sea-water after 10, 20 and 30 minutes. 














er cent. 


per cent. 


• cent. 
































The table shows that eggs segmented in all of the solutions, 
best however in the stronger solutions, although in one case the 
eggs became so injured by the ether that they did not develop 
further than the segmentation stages. In a check series, in which 
the self-fertilized eggs were put into sea-water, about ten per cent, 
of the eggs segmented. There may have been some source of con- 
tamination, or else, and this seems more likely since the indi- 
vidual had been isolated 20 hours, self-fertilization took place 
on a larger scale than usual. 

Experiment IF. Eggs and sperm were mixed in ether-solutions 
of 4, 2, I, 0.5 per cent. 

Ether 2 min. 4 min. 

4. o o ■ «*-,- 

2. 40 I 

I- 35 30 

0.5 5 I 

These results show that the 4 per cent, solution was too strong, 
while the 0.5 per cent, solution appears to have been too weak. 
The injurious action of the 4 per cent, solution appears to have 
been mainly on the sperm rather than on the eggs, for these 
eggs after they had been in sea-water 4 hours were capable of 
being cross-fertilized, and 25 per cent, of them developed. 

Experiment V. Eggs and sperm were put into a 2 and into a 
0.5 per cent, ether-solution and removed after 5 minutes. 


5 mm 





144 T. H. Morgan. 

It Is Interesting to note In this case, In which so large a per- 
centage of the eggs were self-fertUIzed In ether, that of several 
hundred eggs of the same Individual, to which sperm was added, 
but which were kept In sea-water, not one segmented. It was 
also found, and will be referred to again later, that when the 
sperm alone was put into a 2 per cent, soluton, of ether for five 
minutes, and was then added to eggs of the same individual In 
sea-water, 70 per cent, of the eggs segmented. 

Experiment VI. In this experiment the eggs and the sperm 
were put into ether-solutions of 2 and of 0.5 per cent, and 
removed after ten minutes. In one lot the eggs were self-fertil- 
ized, in the other they were cross-fertilized. 

Self-fert. Cross-fert. 

• Ether 10 min. 10 min. 

2. o 100 

0.5 o 100 

In this case although no self-fertihzation took place, all the 
crossed eggs which had also been In the ether-solution developed, 
showing that the solutions have no baneful effect on cross-fertiliza- 
tion. The lack of self-fertilization shows that the sperm were 
not sufficiently acted upon by the ether-solutions employed to effect 

Experiment VII. This experiment shows how slight a differ- 
ence in the conditions may cause great differences In the result. 
The Individuals had been Isolated 48 hours. One lot of self- 
fertilized eggs was kept in water and allowed to stand there 20 
minutes. The eggs with the surrounding sperm were then put 
Into ether-solutions of 2, i and 0.5 per cent, for 20 minutes, and 
then returned to sea-water. None of these segmented. Another 
lot of eggs from this individual were mixed with sperm of the 
same individual and put into a 2 per cent, ether-solution for 
ten minutes and then carried back to sea-water. Here 95 per 
cent, of the eggs segmented. On the other hand some of these 
same eggs taken from the ether after 5 minutes did not divide. 
The following experiments were also carried out with other self- 
fertilized eggs of the same individual. 

Self -Fertilization Induced by Artificial Means. 145 

Ether 15 min. 30 min. 

2. 45 o (only 3 eggs) 

I. 30 40 

Experiment VIII. The eggs and the sperm of one individual 
were mixed in ether-solutions of 4, 3, 2, i, 0.7, 0.5 per cent., and 
removed after 10, 20, 30, 60 minutes to sea-water. It was 
noticed that the spermatozoa were very sluggish in sea-water, and 
although somewhat more active in the ether solutions, yet their 
activity was not marked. Of the eggs, which appeared to be in 
excellent condition, only three segmented, two in the i per cent. 
(20 minutes) and one in the 0.5 per cent. (20 minutes.) 

The eggs that had not segmented after the ether treatment 
were fertilized, after they had stood 6 hours, with sperm from 
another individual, and three quarters of an hour later nearly all 
had divided normally into two cells. It was observed that the 
spermatozoa of this second individual, used for cross-fertilization, 
were also very inactive in their own fluid. They seemed to be 
more active in the extract from the ovary of the first individual. 
This ovary had also stood 6 hours in sea-water. 

At the same time another experiment was made in which the 
eggs of another individual, that had been isolated for 24 hours, 
were put into ether solutions of 3, 2, i per cent, for 10, 20, and 
60 minutes, and then returned to sea-water. None of these eggs 
divided, except one in the 3 per cent, solution (60 minutes). It 
was observed in this case that the sperm was inactive even in the 
ether-solutions, but nevertheless this same sperm, not in ether, 
cross-fertilized the eggs of the other individual in the preceding 
series, and also, as stated, appeared to be somewhat active in 
the extract of the ovary of the other individual. 

Experiment IX. The eggs and sperm together were put into 
ether-solutions of 5, 4, 0.7 per cent. The sperm were active 
even in sea-water. The individual had been isolated for two days. 



:o mm. 

20 mm. 

30 mm. 




(7 out of 8 eggs) 



(9 out of ID eggs) 

146 T. H. Morgan. 

There was another series in this set in which the ether was 
stronger (about one-half saturated). None of the eggs from this 
solution segmented, but they became filled with clear spots. In 
another check series, self-fertilized but kept in sea-water, none 
of the eggs developed. 

Experiment X. Sperm alone was put into ether solutions of 
6, 4, I, 0.5 per cent. It was removed (along with some of the 
surrounding fluid) and added to the eggs after 2 and 10 minutes. 

Ether 2 min. lomin. 

6. o o 

4. o o 

I. 20 10 

0.5 4 50 

It appears from this experiment that it suffices to put only the 
sperm into the ether-solutions to bring about self-fertilization, but 
it should not be overlooked that a certain amount of the ether is 
carried over with the sperm when the latter is added to the eggs. 
The amount, it is true, will be small, since the eggs stand in 
water which further dilutes the ether, but so long as this source 
of error is present, and it is very difficult to remove it entirely, 
the result does not show conclusively that the ether acts on the 
sperm alone, although I think this is the more probable inter- 

A check series of experiments was also made in which both 
eggs and sperm were put into solutions of the same strength as 
those given above, for 15 minutes and then removed to water. 

Ether. 15 min. 

6. o 

4. o 

I. 100 (but only ten eggs present.) 

0.5 90 

It is evident from both of the foregoing tables that only the 
weak solutions were effective, and from the first table it appears 
that this must have been the result of injury to the sperm. It 
can easily be seen that the eggs also are killed in a few minutes 
by a 6 per cent, solution of ether. 

Self -Fertilization Induced by Artificial Means. 147 

Experiment XI. Eggs from the oviduct were put into ether 
solutions of I and 0.5 per cent, for one-half and three-quarters 
of an hour; then washed in a small amount of fresh water and 
fertilized with the sperm from the vas deferens of the same 
individual. The experiment was carried out primarily in order 
to see if the eggs were affected by the solutions, so that they 
could be subsequently self-fertilized, but it is obvious that this 
test is not a good one, since the eggs will carry with them, de- 
spite the partial washing in water, some of the ether which may 
then act on the sperm. The results were as follows : 

Ether ^ hour ^ hour 

i.o o o 

0-5 4 5 

Without a check series, which unfortunately was not made, it 
is difficult to decide whether the small number of eggs that were 
self-fertilized was due to the action of the ether on the eggs 
or on the sperm. The experiment must be repeated on a more 
elaborate scale. 

Experiment XII. In this experiment with two individuals, 
weaker solutions of ether were used. In one lot the sperm alone 
was put into ether, and then added to the eggs. In the other 
lot both eggs and sperm were put together into the ether. I 
omitted recording the time in the ether, but it was probably about 
five minutes. 


Sperm and Eggs in Ether 

Sperm only in Ether 






Sperm and Eggs in Ether 

Sperm only in Ether 








This experiment shows that the sperm of the first individual 
was incapable of self-fertilization, even with the ether present. 
In the other individual, the sperm was good, and there was a 

148 T. H. Morgan. 

great deal of it; hence, no doubt, the excellent results in the first 
column. What is especially significant is that the best results 
were obtained when the eggs and the sperm were put at the same 
time into the solution together. This may mean the ether has 
some effect on the eggs as well as on the sperm, or that the most 
effective period of activity for the sperm is immediately after 
it comes into contact with the ether. My experiments do not 
sufl^ce to settle this point, but that the spermatozoa are still 
capable of cross-fertilizing, after they have been in the ether 
for some time, is shown by the following result. After four hours 
the eggs of the first individual were mixed with the eggs and the 
sperm of the second individual. Later it was found that all 
the unsegmented eggs had been fertilized. The ether had no 
doubt largely evaporated. 

The preceding twelve experiments with ether-solutions gave 
definite results, although in a few cases the number of eggs self- 
fertilized was small. It should be stated that there were ten 
other individuals in which self-fertilization in ether did not take 
place. This does not detract, I think, from the value of the 
successful experiments, because, as has been shown, the sperm 
is sometimes incapable of fertilizing even the eggs of another 
individual. The following experiments were carried out in order 
to examine this question further. It will be observed that parallel 
experiments with ether were also performed. 

In each series five individuals were used. The eggs of each 
were fertilized with the sperm of every other individual. The 
following scheme shows the order in which the eggs were crossed. 
An individual having been opened, the eggs were removed from 
its oviduct and distributed in five dishes, A-A. Another individual 
was then opened (using, of course, different scissors, pipettes, etc.) 
and Its eggs distributed to the next line of dishes, B-B. The same 
method was followed for the other three individuals. The sperm, 
a, of the first individual was then taken out and put into a small 
amount of water. It was then distributed to one set of eggs from 
each of the other Individuals, B, C, D, E; then the sperm of B 
was taken out and applied to another set of eggs. The process 
was repeated until all the eggs were supplied with sperm. The 

Self -Fertilization Induced by Artificial Means. 149 

sperm in each case is indicated in the table by the small letter 
used as an exponent. The first set of A-eggs was as a rule fer- 
tilized with the e-sperm and the last set of E-eggs with the a-sperm. 
Experiment XIII. — 

&100 EnOO E'^lOO EnOO E^' 85 


O 98 0100 C 75 O 75 OlOO 

B^ 99 BaOO B-^lOO B<^ 100 B'^ 100 

A^(adtsptm) ^-'G'niil) A<^100 A"^ 85 A^ 99 

In this experiment practically all of the eggs were fertilized 
by the sperm of another individual. When fewer than the total 
number segmented (fertilized), immature eggs may have been 
present. As a check series A, B, C, D and E were self-fertilized. 
None segmented, except in E, where two eggs out of the twenty 
present, i. e., 10 per cent, divided. 

The two ether series (self-fertilized) of these same eggs gave 
the following results : 















Experiment XIV. An experiment similar to the last was carried 
out, with five other individuals, and gave the following results: 



E^ 30 

En 00 

E^ 4 



D^ c:.) 

■P)e /no npe\ 
^-^ V eggs ) 




pd / no V 






B'^ 50 







Despite a few slight discrepancies in this table, the main result 
is clear. In only two of the five individuals was the sperm capable 
of cross-fertilization, namely, the e-sperm and the d-sperm. 

There was also a self-fertilized series of these eggs, and in this 
not any of the eggs segmented. The ether series gave the fol- 
lowing results : 

T. H. Morgan 




o o o 


? 2 


Ether A B C D E 

0.5 00 o o 10 (only ten eggs.) 

i.o o o ? 2 50 

It becomes evident from this result that the frequent failure 
of the sperm and eggs (mixed together) to self-fertilize in ether 
is due to the poor quality of the sperm. The poor sperm does 
not cross-fertilize, and presumably for the same cause it can not 
always be made to self-fertilize even in the ether. That sperm 
that is too poor to cross-fertilize may sometimes self-fertilize with 
the help of ether I hold to be possible. I regret that I did not at- 
tempt to determine whether poor sperm, that will not cross-fer- 
tilize, can be made to do so by means of ether, but other experi- 
ments lead me to think that it would often do so. 

Experiment XV. In the following experiment all of the sperm 
appears to have been good except that of A, whose eggs, how- 
ever, were in excellent condition. 



E" 85 

E^ 90 

£<• (^;^ 





D^ 20 

D^ 90 




C" 100 



C^ 90 



B^ 100 



B^ 100 


A" 70 




It is clear that the a-sperm was poor, although it did well 
In C"" in which 70 per cent, of the eggs divided. 

There was also a self-fertilized series in which none of the 
eggs segmented, except 5 per cent. In B. (In C, 90 per cent, of 
the eggs divided, but this may have been due to accidental con- 
tamination.) In the ether series the following results were ob- 














In this case although the spermatozoa of B, C, D, E were capa- 
ble of crossing, they self-fertilized In ether very poorly, except In 
C, where good results followed. 

Experiment XVI. In this experiment again only the first in- 








12 Utls) 



Self -Fertilization Induced by Artificial Means. 151 

dividual produced poor sperm, yet It did fairly well In one case, 
and In ether gave some results. 

E^i 5 ^. .^JE^IOOO^^E^ 80:;':^^ E" 100 E^ 70 

D- OriH-'DnOO^r D'^ lO(Ss) D^lOO D^ 30 

O; [^ 0100 0100 0100 C 100 

B^ 25Gtj,) B^c-) B^G-) B^ (,- ) B^ ( -j 

A^ 75 AMOO A<^100 A" 100 A^OO 

In the self-fertilized series no eggs segmented. The ether 
series gave the following results : 

Ether A 

0.5 2 

1.0 50 (;Ss) 

The experiments recorded in Exp. XIII to XVI show that 
the sperm is at fault when cross-fertilization does not take place. 
In fact, eggs in the oviduct seem always to be capable of cross- 
fertilization. It is also evident that it Is more difficult to get 
results with ether when the sperm does not cross-fertilize well, 
than when it does act well In this way. From this It seems to me 
very probable that when the ether fails to bring about self- fer- 
tilization the fault lies with the sperm. We may perhaps even 
go further and conclude that the action of the ether in bringing 
about the self-fertilization Is on the sperm alone, but I am not 
In position to prove positively that the action of the ether on the 
eggs may not also enter Into the result. 

In concluding my account of these experiments on Clona, I 
should like to point out that I had constantly in mind the possi- 
bility that the ether might produce parthenogenetic segmentation, 
and that the sperm had in reality nothing to do with the result. 
It was abundantly shown, however, that this was not the case, 
and in the few experiments in which I put this view to the test, 
by keeping eggs without sperm In ether - solutions of various 
strengths, I got no results when the eggs were returned to water. 
It should be noted in this connection that Lyon^ has recently 

1 American Journal of Physiology, IX, July, 1903. 

152 T. H. Morgan. 

recorded that he was unable to cause artificial parthenogenesis In 
Ciona intestlnalls at Naples by any of the ordinary means that 
excite this development in other eggs. 

I shall discuss later the view as to whether eggs may be entered 
by the sperm of the same individual, but fail to develop unless 
incited to do so by some external agent. 

It has been pointed out in the preceding pages that the eggs of 
Ciona may be fertilized after they have been in sea-water several 
hours. I made a test of this again in the following experiment: 

Experiment XVII. Some eggs were cross-fertilized at once, 
others after 30, 80, 125 minutes, with fresh sperm from the same 
Individual. All the eggs developed. A striking fact was ob- 
served in this case. The eggs fertilized late began to segment 
after a shorter interval than did those fertilized at once, so that at 
the 32-cell stage those fertilized last were only one division behind 
the first set, and no doubt soon caught up. It appears that a 
ripening process goes on in the egg as it stands in the sea-water, 
so that it begins to segment more quickly after it is fertilized 
than does an egg fertilized as soon as removed from the oviducts. 
It even appeared that after the first cleavage the rhythm of divi- 
sion was quicker in the eggs whose fertilization had been de- 
layed, but this point needs a special examination which I have not 
yet made. The discovery is all the more significant because the 
first polar spindle Is already formed in Ciona while the egg is in 
the oviduct, and the spindle remains resting in the equatorial plate 
stage until the egg is fertilized; hence the difference In time of 
segmentation can not be accounted for by the time required for 
the breaking down of the egg-nucleus and for the formation of 
the polar spindle after the egg has been removed from the ani- 
mal. Some change must take place In the sea-water, which, while 
it does not cause the polar spindle to pursue Its development, yet 
causes the developments that take place after the spermatozoon 
enters to go on more rapidly. 


The ovaries of Cynthia extend far forward, and have a very 
short oviduct. Each ovary — there appear to be two in each In- 

Self -Fertilization Induced by Artificial Means. 153 

dividual — is double, the halves being united at the distal end. 
Owing to the close proximity of the ovary to the surrounding 
tubes of the testis, it is possible only by very careful manipulation 
to get the eggs out of the cavity of the ovary without cutting into 
the testicular tubes. When it was necessary to separate the eggs 
from the sperm of the same individual, I have carried out this 
operation, but in general the ovaries and the testes were cut up 

For the purpose of studying the effects of self-fertilization 
Cynthia is in many respects inferior to Ciona because self-fer- 
tilization takes place to a very large extent. On the other hand, 
if check experiments are use'd for each individual, this factor can 
be estimated, and the very fact that Cynthia does self-fertilize 
its own eggs to such an extent gives an opportunity to examine 
other aspects of the problem. A much more serious difficulty 
is met with in that artificial cross-fertilization is often unsuccessful 
in this species. Even when the eggs and sperm from a large 
number of individuals are mixed together, fertilization may not 
take place; but in curious contrast to this result are the following 
observations on the egg-laying processes of this animal kept in 
aquaria. On several occasions a number of individuals were put 
together in the same dish. About 5 o'clock in the afternoon one 
after another began to send out jets of eggs and of sperm pro- 
ducing the effect of a lively cannonading. Under these circum- 
stances it was found that every single egg was fertilized. Perhaps 
only ripe individuals sent out their eggs and sperm, or perhaps the 
eggs were mature in all individuals, and the sperm from one or 
two individuals may have sufficed to fertilize all of the eggs. In 
general it is, I think, the sperm of Cynthia that is not good. 
Certainly the spermatozoa are often very sluggish when taken 
from the testis and put into water. May it not be possible that 
when the eggs are laid, Cynthia secretes some other fluid that 
makes the sperm active? This point needs further investigation. 

The best means that I found to determine the extent to which 
self-fertilization of the eggs of Cynthia may take place was to 
isolate some of the individuals early in the day, and observe in 
those that emitted eggs and sperm in the late afternoon the per- 

154 T. H. Morgan. 

centage of eggs that segmented. The following four records 
were obtained in this way: For August ii — 33, 10, 100, 95, 95, 
75, 10 per cent. For August 16 — 30, 30, 10, i, 75, 90, 85 per 
cent. For August 19 — 33, o, 10, 4, 4, o. For August 20 — 12, 4. 
A much larger number of individuals gave off neither eggs 
nor sperm, and some produced sperm and no eggs, and vice versa. 
The results in the above list show all conditions from perfect 
self-fertility to absolute self-sterility, although some of the latter 
cases may have been due to no sperm being given off. 

A few preliminary trials were made with two (A and B), and 
with three (A, B, and C) individuals. The scheme of crossing 
is given in the following diagrams: 
For Two Individuals. For Three Individuals. 














A few examples 

of the results with two Individuals are as 

follows : 


B" A^ 


A" few 

B^ 15 A" 10 



B^ Outln.) A^ 

B^ 00 

A" few 

B^ 20 A" few 
A^ B^OC^^) 
A^ 50 B^ 


Comparing the self-fertilized eggs with the crossed-eggs. It 
Is clear that while self-fertilization did not take place in nine 
cases, and In only one egg in the other case, yet cross-fertilization 
more frequently occurred, but never so completely as when many 
individuals normally deposited their eggs and sperm together. 
In addition to these cases there were three others in which none 
of the eggs, neither self- nor cross-fertilized, segmented. One 
of the results with three individuals is given In the next table : 

A^ B*^ 2 C^ 

B^ few A'' very few B"^ 25 

75 Orare A^ 4 



A^ 50 CIT) 

B^ 1 

C" very few 


Self -Fertilization Induced by Artificial Means. 155 

In this experiment the a- and c-sperm did not self-fertilize, but 
the former did well with C- and the latter with B-eggs. The 
b-sperm self-fertilized to a slight extent, but did no better with 
the A- and with the C-eggs. 

In the next series the results are more striking: 


Here none of the sperm self-fertilized the eggs. The a-sperm 
did quite well with the B- and C-eggs (95 and 50 per cent) . The 
b-sperm did well with the A-eggs, but not with the C-eggs. The 
c-sperm did well with the A-eggs, but not with the B-eggs. It 
may appear from the preceding table that there is something more 
involved than simply the question of good sperm, for the same 
sperm appears to act differently with different eggs. 

Another experiment with three individuals gave no eggs self- 
fertilized, but good cross-fertilizations with the c-sperm; less good 
with the b-sperm. These experiments should be carried out on 
a larger scale, and at different times of the year, but they suffice 
to show that self-fertilization is very infrequent when the process 
is an artificial one. It takes place to a considerable extent in some 
cases when eggs are normally laid. Moreover the artificially 
crossed eggs do not segment nearly so well in Cynthia as in Ciona. 

The next experiment shows the action of ether on self- and 
cross-fertilized eggs. Some of the eggs and sperm of one in- 
dividual, A, were removed and put into sea-water. Other eggs, 
A% were self-fertilized in an ether-solution, and a third lot, A^, 
were crossed with sperm from B (A-sperm was also present). 
The same process was carried out with B which was crossed with 
sperm from A. 


A^' few B'' few 

A'' very few B^ very few 

The results show that the self-fertilized eggs in ether did as 
well as those that were crossed, but none of the eggs in water alone, 

156 T. H. Morgan. 

with their own sperm, segmented. Another similar experiment 
with two other individuals gave the following results : 


A^' B" 

A" 50 B' r^'^^T) 

In another set the ether appears to have been too strong, yet 
50 per cent, of A'' divided. 

In another experiment, 10 per cent, of the self-fertilized eggs 
in ether segmented, and 50 per cent, of the crossed. 

In another, 5 per cent, of the self-fertilized eggs in ether seg- 
mented, and 75 of the crossed. 

The next set is more instructive : 


A^' 100 B*^ 

A^ 2 B^ 4 

It is clear that the ether had a marked effect in A"*, making all 
of the eggs self-fertilize. This is all the more interesting be- 
cause none of the eggs without ether self-fertilized. Both eggs 
and sperm of the B- set appear to have been in poor condition, so 
that the sperm did not cross-fertilize, or the eggs become cross- 
fertilized, to any extent. 

In searching for other substances that might act on the sperma- 
tozoa as does tht ether, I tried, amongst other things, a solution 
of ammonia in sea-water, and this I found made the spermatozoa 
even more active than the ether. Dilute solutions of alcohol 
from I to 10 per cent, also excite the spermatozoa to greater ac- 
tivity. Certain salt-solutions, ammonium chloride (i, ^, ^ per 
cent.), magnesium chloride (2 per cent.), and sodium chloride (i 
per cent.) appeared also to act on the sperm, but much less ef- 
fectively than does ether, alcohol, or ammonia. In the alcohol 
series of i, 3, 5, 6, 10 per cent., it was found that i per cent, made 
the sperm very little more active; 3 per cent, more so; 5 per cent, 
most active; 6 per cent, less; 8 per cent, no effect; 10 per cent., no 
effect. The last two solutions undoubtedly injured the sperm. 
In another series, 7 per cent, gave the best results. 

Self-Fertilization Induced by Artifidal Means. 157 

A few experiments were carried out in order to see if the sperm 
made active by the alcohol, would self-fertilize the eggs when 
it would not do so without the stimulus. Here, as in the preceding 
series, the same lettering will be used in the tables. A^ self- 
fertilized in sea-water, A^ self-fertilized in alcohol-solution, A'' 
crossed in alcohol-solution. 

A^ B'^O 

A^ [Alcohol] few B" [Alcohol] several 

A" 20 B^ 50 

In this experiment while no eggs were self-fertilized in sea- 
water, a few or several (the percentages were not recorded) were 
self-fertilized in alcohol, but even more developed in the crossed 

In another experiment only one individual was used. The 
eggs, self-fertilized in sea-water, did not segment, but 10 per cent, 
did so in a 3 per cent, soluton of alcohol, and 50 per cent, in a 5 
per cent, solution of alcohol. 

Solutions of ammonia gave similar results. Sperm and eggs 
were mixed together in very dilute solutions of ammonia. Many 
eggs divided and of these most appeared, from their method of 
division into several cells at once, to be polyspermic. Some of 
the sperm from the last lot was added to eggs in sea-water. 
Fewer eggs were fertilized, but several that were fertilized were 
polyspermic. Eggs (not separated from their own sperm) were 
crossed in ether. All of these were polyspermic. Another set 
gave almost identical results. 

It is clear from these experiments that those solutions that 
make the spermatozoa more active often induce fertilization of 
the eggs, when such a fertilization does not take place without 
the use of the solutions. The activity of the sperm and the fer- 
tilization of the egg appear to be directly connected. This point 
will be more fully discussed later. 

On each side of the body of Molgula there is an ovary sur- 
rounded by a testis. It is very easy to open the central cavity 
of the ovary, and remove the eggs without cutting the testis. 

158 T.H.Morgan. 

A few preliminary experiments showed that the sperm of Mol- 
gula fertilizes the eggs of the same individual. The -following 
illustrations will show the great powers of self-fertilization of 
this species: 

A^ 85 
A"^ 90 

B" 90 
B^ 100 


A" 100 

B" 2 

B^" 90 [irregular 

A=^ 100 

B^ 100 
B^ 100 

A" few 


A^ 90 
A" 100 



These cases make it clear that the sperm is capable of fer- 
tilizing the eggs of the same individual. Whether the sperm 
of another individual is prepotent I did not attempt to determine. 
There were only a few cases in which neither self- nor cross- 
fertilization was effective, and whenever good crossing was accom- 
plished self-fertilization was also realized, showing that when the 
sperm is good, it will readily fertilize the eggs of the same 
individual. Since similar results were obtained when three in- 
dividuals were used it will not be necessary to give the latter 
cases. The experiments were not extensive enough to show whether 
good sperm affects the eggs of certain individuals better than it 
does others, but Molgula is not well suited to test this point. 

It occurred to me as possible that in Cynthia and in Molgula 
the power to self-fertilize the eggs might be due to the eggs 
coming from the ovary on one side of the body, and the sperm 
from the other side. Conversely, if this were true, the lack of 
self-fertilization in Ciona might be connected with the presence 
of only one ovo-testis. I examined this possibility for Molgula. 
The eggs from the small ovo-testis were fertilized with sperm 
from the same side, and other eggs with the sperm from the other 
side. In both cases all the eggs were fertilized. Conversely, the 
eggs from the large ovary were fertilized with sperm from the 

1 In B the sperm was probably bad. The Ab must therefore have been self- 
fertilized. The same conditions hold also for the second couple. 

Self -Fertilization Induced by Artificial Means. 159 

same side, and others with sperm from the opposite side. Here 
also all the eggs segmented. It is perfectly evident, therefore, 
that the question of self-fertilization in Molgula is not connected 
with the double condition of the ovo-testis. 


In order to find out how generally ether, alcohol, and ammonia 
excite to greater activity the movements of cilia, of flagella, and 
of the spermatozoa of other animals, I made a few experiments 
on certain protozoa and on the spermatozoa of the frog and of 
the rat. 

Ether, 5 per cent, stops the movements of paramoecium, and 
kills stentor; 3 per cent, slows up the movements of the former, 
and causes stentor to throw off its outer layer; the movements of 
free swimming v^orticellae seemed to be increased; 2 and i per 
cent, hasten the movements of paramoecium and of stentor. 

Alcohol of 6 and of 8 per cent, slow down the movements of 
paramoecium and stylonychia, and cause stentor to disintegrate; 
10 per cent, kills; 4 per cent, appears to be near the limit, and 
seems to increase their activity; 2 per cent, clearly increases their 

Ammonia 1/200 per cent, kills paramoecium, stentor, and sty- 
lonychia; and even 1/2000 also kills; 1/5000 per cent, seems to 
make these protozoa somewhat more active, but I have not suf- 
ficiently tested this solution. 

Some of the same solutions were used with euglena, which 
moves by means of an anteriorly directed flagellum. Ether 5 
per cent, makes them somewhat more active; 3 per cent, less so, 
and 2 per cent, gives no very noticeable effect. Alcohol 10 per 
cent, kills; 8, 6, and 4 per cent, make them swim more actively; 
2 and I per cent, give no definite result. Ammonia 1/200 kills; 
1/2000 per cent, does not appear to make euglena more active, 
but other strengths should be tried. 

A male spotted frog [Rana halecina) was killed in No- 
vember; its testes opened, and the immobile sperm squeezed 
out into normal salt-solution. It was found that it took 
some minutes to get a noticeable effect. Ether 5 and 2 

i6o T. H. Morgan. 

per cent, caused the spermatozoa to show some movement 
in the course of 15 minutes. Alcohol gave better results. A 
10 and a 6 per cent, solution awakened the spermatozoa to ac- 
tivity; a 4 per cent, gave the best results of all. In no case, how- 
ever, was the activity very great. No movements were detected 
in ammonia-solutions, but only two strengths were used. 

These scattering and incomplete observations show that these 
substances are in all probability general stimulants for protoplas- 
mic activity of certain kinds. * 

I have also made a few experiments with the spermatozoa 
of mice. The spermatozoa were taken directly from the testis 
of a mouse that had just been killed. The solutions were added 
to a drop of the sperm squeezed out from the testis into a drop 
of physiological salt-solution, consequently the dilution is greater 
than actually given by the percentage. In certain strengths of 
ether (5 per cent.) and of alcohol (8 per cent.) it appeared that 
the movement was increased; with ammonia I did not get satis- 
factory results. The observations are made more uncertain here 
because, when the testes are opened, spermatozoa in all stages 
of development are found, and are consequently acted upon dif- 
ferently by the solutions. It would be more satisfactory to use 
a larger animal and take the spermatozoa from the vasa deferentia, 
where they arc all fully formed. It is certain, however, that al- 
cohol and ether do not produce as great effects on these sperma- 
tozoa as they do on the spermatozoa of the ascidians and of 
some other marine animals that I have examined. 

In one of the preparations of the mouse testis the water began 
to run out at one side and it became apparent at once that the 
spermatozoa all turned and headed up-stream. It has been re- 
corded by Kraft that spermatozoa swim in the opposite direction 
to that in which the cilia of the oviducts act. My observation 
suggests that movement in this direction is not due to the sper- 
matozoa swimming against the direction of the greatest action 
of the cilia, but against the stream that is produced by the cilia. 
The movement may be a simple physical phenomenon — the lighter 
tails of the spermatozoa being swept backwards by the current 
so that the heads are turned up-stream, and the contraction of 
the tail then causes the spermatozoon to travel in this direction. 

Self -Fertilization Induced hy Artificial Means. i6i 

In later experiments the sperm was taken from the vasa defer- 
entia, and put first into distilled water where the spermatozoa 
remained quiescent. If a drop of salt-solution (water loo, NaCl 
0.75) was added to a drop of water containing the spermatozoa, 
they became active in the course of a minute or less, and their 
activity continued to increase for several minutes longer, when 
they remained active for some time. If a drop of a 5 per cent, 
ether solution is added to a drop of water containing quiescent 
spermatozoa, no result is seen at first, but after ten minutes I 
have observed a slight vibration of the spermatozoa. If now 
after the ether has been added, a drop of the salt-solution is also 
added,the spermatozoa become active, but it is difficult to determine 
whether they become more active than when the salt-solution alone 
is present. Certainly there is no marked difference. If a drop 
of 8 per cent, alcohol is added to a drop of water containing the 
spermatozoa no activity is observable, but if then a drop of salt- 
solution is also added the spermatozoa begin to swim, showing 
that the alcohol had not injured them, although it had failed to 
arouse them to activity. Several strengths of KOH (3 per cent, 
and weaker) were tried, but without effect; yet if salt-solution was 
added later some slight activity was seen. 

In another series of experiments the spermatozoa quiescent in 
water were first made active by adding the salt solution. If ether 
was then added no decided effect on the sperm could be seen 
when their activity was compared with that of check preparations 
of salt-solution only. It appeared sometimes as though the ether 
did make the activity more pronounced, and the movement of the 
spermatozoa appeared somewhat different in the two cases. In 
the ether the motion was more jerky, and in the salt solution more 
sinuous and normal. 

The following. solutions were also tried: The sperm was first 
put into a drop of water, and then a drop of the solution was 
added, NaHCO^ 0.625 per cent, caused the sperm to vibrate 
rather actively; NasCOg, 5.0 per cent, caused a little activity after 
five minutes; KCl, 0.5 per cent, caused greater activity than 
did the sodium carbonate, while CaCl caused somewhat less vibra- 

1 62 T. H. Morgan. 

These, and some other experiments that need not be described 
here, show that salt-solutions of various kinds have a marked 
effect in arousing to activity the inactive spermatozoa of the vasa 
deferentia. They also make active, spermatozoa that are quies- 
cent in distilled water. On the other hand ether, alcohol, and 
ammonia, which proved so efficient for the spermatozoa of the sea- 
urchin and starfish, appear to have little effect on the spermatozoa 
of the mouse. 

The more fundamental physiological question as to the nature 
of the action of these different substances I shall not attempt to 
discuss without a further basis of observation and experiment to 
go upon. Enough has been seen, however, to suggest that the 
substances act as a "stimulus," which is perhaps not dissimilar in 
kind from that which causes some eggs to begin to develop, or 
a nerve impulse to start, or a muscle to contract. Here also we 
may urge, as I have urged elsewhere^ in opposition to Loeb's 
conclusion in regard to the action of certain agents in causing 
artificial parthenogenesis, that the nature of the stimulus is of 
such a kind that the result depends much more on the structure 
or the composition of the living thing than upon the kind of 
stimulus employed. So unstable is the living organization that 
the sHghtest change brought about in it by chemical or by physical 
means suffices to set into action a perfectly definite and pre- 
arranged series of events. 


The action of ether, ammonia and alcohol on the speramtozoa 
of Ciona, arousing them to greater activity and thus, under certain 
conditions, bringing about the fertilization of the egg, raises the 
question as to whether in the higher animals a similar action may 
not result from the application of these and of other substances, 
and also whether the secretions of some of the glands connected 
with the reproductive system may not have a similar effect on the 

When I tried to find some substances that might bring about 
self-fertilization in Ciona I was not aware that there had already 

1 Science. N. S. XI. 1900. Pp. 178-180. 

Self -Fertilization Induced by Artificial Means, 163 

been made several experiments on the action of solutions on the 
spermatozoa of other animals. I find that there are quite a num- 
ber of observations of this sort, although none of the observers 
have had in view the same question with which I was especially 

Kolliker in 1856 carried out an extensive series of experiments 
on the effect of different solutions on the spermatozoa of the 
bull, dog, rabbit, horse, and also made a few observations on the 
spermatozoa from a human cadaver. He found that water alone 
quickly brings spermatozoa to rest, but does not kill them. They 
can be aroused to activity by adding, for instance, a 10 per cent, 
solution of disodium phosphate.^ Many other substances were 
found favorable to the activity of the spermatozoa, such as blood- 
serum, sugar in certain strengths, sodium chloride, caustic pot- 
ash, etc. 

The caustic alkalies (potassium, sodium, and ammonium hy- 
droxide) were found to be especially powerful excitants. Kol- 
liker also tried a number of other solutions, such as three different 
kinds of sugar, glycerine, gum, etc., which in certain strengths 
cause increased activity; also urea, gall, morphine, strychnine 
(nitricum), which have an indifferent effect. He also tried al- 
cohol, creosote, chloroform, ether, alkaloids, and tannin, which 
have an injurious effect. Kolliker also examined the action of the 
secretions of the glands of the male reproductive organs — the 
uterus masculinus, prostate and Cowper's glands. He found 
that these secretions excite the spermatozoa to greater activity.^ 

The much more recent experiments of Steinach bear even more 
directly on the present problem. He found that after removal of 
the glandulae vesiculares ("receptaculum seminalis" of some writ- 
ers) of the male rat, that, although the sexual instinct remained, 
the number of young that were born was much decreased. When 

1 Kolliker gives the formula 2Na0H0P0g, which is no doubt disodium 
phosphate, now written Na HPO . 

2 Moleschott and Richetti are quoted by Kolliker as recognizing the favorable 
action of sodium salts on the spermatozoa. Quatrefages found that the sper- 
matozoa of the weasel showed a "surexcitation" in 64 parts sea water to one 
part sea salt. Newport found that potassium carbonate, and also 1/480 of 
potassium salt made the activity of the spermatozoa of the frog greater. 

164 T. H. Morgan. 

this gland, as well as the prostate, was removed no young at all 
were born, although frequent union with the females took place. 
The results may be due to the semen being insufficiently diluted 
when it is not mixed with the secretions of the glands, or else to 
the absence of proper excitation of the spermatozoa when the 
gland-secretion is removed. That the spermatozoa may be nor- 
mally acted upon by the secretion of the glands was shown by 
Steinach in the following way: Sperm from the vas deferens was 
mixed with a physiological salt-solution. A drop was placed 
under a cover slip and the edges sealed to prevent evaporation. 
The preparation was kept at a temperature of 35° to 37° C. A 
similar preparation was made with the secretion of the prostate. 
In the former the spermatozoa began to lose their activity in one 
and a half hours, and after three hours had come completely 
to rest. In the other preparation, that containing the extract 
from the prostate gland, the spermatozoa were active after 1 1 
hours, and ceased to move altogether only after 22 hours. This 
experiment shows that the secretion of the gland prolongs greatly 
the period of activity of the spermatozoa. Whether it excites 
them to greater activity is not stated, but Kolliker's results leave 
no doubt on this score. The decrease in the fertilizing power 
when the glands were removed may well be connected, as sug- 
gested above, with the lessened activity of the spermatozoa. 

Duller has recently studied the question as^ to whether the 
spermatozoa of the sea urchin are attracted to the egg, — in other 
words, whether, as some authors have assumed off-hand to be the 
case, there is a chemotactic action of the egg on the spermatozoa. 
He points out that although Strasburger claimed that the egg 
of Fucus excretes a substance that attracts the spermatozoon from 
a distance of two diameters of the egg, Bordet and Buller himself 
have failed to confirm this statement. Massart thinks that In 
the case of the frog the meeting of the spermatozoon and the egg 
Is purely accidental. Buller finds for the sea-urchins, Arbacia, 
Echinus, and others, that when the spermatozoa are set free near 
the egg they show no tendency to swim towards it. The dense 
collection of spermatozoa that forms around the egg is due to 
those that happened to run into the jelly sticking there. These 

Self-Fertilization Induced by Artificial Means. 165 

spermatozoa then proceed to bore into the jelly; most of them 
in a radial direction, although a few can be seen to go in obliquely, 
or tangentially. The same phenomenon occurs in unripe eggs, 
and in eggs that have been killed in weak osmic acid and the 
acid washed out. It is improbable, therefore, that chemotaxis 
has anything to do with the result. 

In order to see if any substance is given off by the eggs that 
attracts the spermatozoon, the eggs were taken from the ovary, 
carefully washed, and allowed to stand for 2 to 12 hours in a 
small amount of sea-water. Capillary tubes were then filled with 
this water and placed in a drop containing spermatozoa. The 
spermatozoa did not show any tendency to collect around the 
openings of the tubes. Several other substances were tried in 
the tubes in the same way, — salts, sugar, ferments, acids, alcohol, 
etc. — but no chemotaxis was discovered. 

The spermatozoa of the sea-urchins swim in spirals. Coming 
into contact with a surface, the spiral is changed to a circular 
movement due to contact. Buller considers whether the radial 
path taken by most of the spermatozoa after they have entered 
the jelly is due to stereotropism. He reaches the conclusion that 
while theoretically this assumption will explain the phenomenon, 
yet conclusive evidence in favour of this view is lacking. He 
suggests that it may be possible to find a purely physical solution 
of the problem. 

Von Dungern has examined the question of cross-fertilization 
from the point of view of the different substances contained in the 
egg, and has reached some conclusions of the greatest interest. 
He finds that the egg of the starfish, Jsterias glacialis, contains a 
substance that acts as a poison on the sperm of the sea-urchin 
(Echinus or Sphaerechinus) . The minimal lethal dose for the 
sperm mixed with 2 ccm of sea-water varies considerably with the 
individual; for Echinus between 1/800 to 1/6400 part is fatal 
in half an hour. Von Dungern tried to obtain an antitoxin from 
the blood of the rabbit that would neutralize the effect of the 
poison of the eggs, hoping that it might be possible in this way 
to bring about the cross-fertilization of the egg of the starfish 
by the spermatozoon of the sea-urchin. He found, however, that 

1 66 T. H. Morgan. 

the serum of the normal rabbit already contains a substance that 
has a powerful antitoxic action on the poison of the starfish, so 
that it was not necessary to obtain an antitoxin by injecting the 
poison into the rabbit. The antitoxin of the rabbit's serum was 
added to water containing the eggs of Asterias, and then sperm 
from a sea-urchin was supplied. Von Dungern often obtained 
two- and four-cell stages in this way, but the results were uncertain, 
and he could not decide whether fertilization had or had not 
taken place. It seems not improbable, I think, that the outcome 
may have been due to artificial parthenogenesis which occurs 
very readily in the eggs of certain starfish; in fact, it is very 
difficult to prevent its occurrence, unless the eggs are very care- 
fully handled. 

The same poison that is present in the eggs of the starfish is 
also secreted by the skin. It is also rendered harmless by the 
rabbit's serum. In the sea-urchin there is a poisonous substance 
in the gemmiform pedicellariae, which is very injurious to the 
sperm of the starfish. If lOO of the pedicellariae of Sphaerechi- 
nus are rubbed up in one ccm of sea-water, the solution will de- 
stroy in a quarter of an hour the sperm contained in ten to tw^enty 
litres of sea-water. The minimal lethal dose for 2 ccm is 1/5 120 
to 1/ 1 6240 ccm. The spermatozoa of Sphaerechinus itself are 
killed by this fluid, but a much stronger dose is necessary. On 
the other hand an extract of the egg of Echinus, Sphaerechinus, 
Strongylocentrotus, or Arbacia does not kill the spermatozoa of 
the starfish even in the strongest solutions. What then prevents 
the spermatozoa of the starfish from entering the eggs of these 
sea-urchins? There Is another factor. Von Dungern thinks, that 
interferes with this combination. The egg membrane of these 
urchins has an agglutinizing effect on the spermatozoa of the 
starfish. This agglutinizing effect appears to be the same phe- 
nomenon as that seen "whenever cells of any kind are introduced 
into the body of another animal." So far as this process is in- 
volved in the union of germ-cells. Von Dungern thinks that under 
certain conditions it might assist the fusion, while under others 
it might interfere with it. Thus two naked and equivalent cells 
might be helped to unite, while an egg surrounded by an agglu- 

Self -Fertilization Induced by Artificial Means. 167 

tinizing jelly would fail to be fertilized. The substance in the 
sea-urchin's egg that agglutinizes the starfish sperm can be ren- 
dered ineffective by the rabbit's serum. Not all starfish sperma- 
tozoa are agglutinized by the jelly or by the egg-substance of all 
the different sea-urchins. In Sphaerechinus it fails to occur. 
Therefore in this case the failure to cross-fertilize must be due 
to some other factor, and, in fact. Von Dungern claims to have 
found still another substance in the sea-urchin's egg that excites 
to greater activity the immature and quiescent spermatozoa of 
the starfish. These immature sperm, made active by this sub- 
stance, are then capable of fertilizing the eggs of the starfish. 
He found that weak doses of chloral hydrate and of cocaine also 
make these quiescent spermatozoa active, and that rabbit's serum 
has a marked effect. Von Dungern believes further that these ex- 
citing substances may actually prevent, in certain cases, the cross- 
fertilization, because they may change the kind of reaction shown 
by the sperm. He observed that the spermatozoa of those 
species that do not normally show rotational movements when 
they come in contact with surfaces, usually do so when the excit- 
ing substances just mentioned are present. It does not appear 
to me, however, that this is an altogether satisfactory explanation 
of the failure of cross-fertilization in these cases. 

Von Dungern also examined the question as to whether the 
egg secretes a substance that favours fertilization by its own 
sperm. He believes that he has also discovered such a substance. 
The eggs of Echinus (or of Sphaerechinus) are rubbed up and 
mixed with pieces of jelly that have been carefully washed. When 
sperm is added to the water in which such pieces lie they stand 
vertically to the surfaces of the pieces. If on the other hand the 
pieces of jelly are not mixed with the substance from the egg, the 
spermatozoa simply rotate on the surface of the jelly, and do not 
stand vertically. Starfish spermatozoa with Arbacia jelly be- 
have as with simple jelly alone, i. e.^ they do not stand vertically. 
The vertical position of the spermatozoa is due. Von Dungern 
thinks, to the presence of some substance in the extract that lowers 
the excitability of the spermatozoon to contact, and hence it takes 
a vertical position. He also points out that this same substance 

1 68 T. H. Morgan. 

causes the spermatozoa to lose their power of movement in a 
short time. Thus, while Von Dungern finds no evidence of a 
substance in the egg that attracts the sperm, he believes that there 
may be present in some eggs a substance that favours the fer- 
tilization of the egg, by causing the spermatozoon to assume that 
position in the jelly that is most likely to bring them to the surface 
of the egg. 

Loew has attempted to show by an experiment, which is not, I 
think, well suited to prove his point, that the spermatozoa of the 
rat are attracted to, i. e., that they are positively chemotactic to, 
the slime layer of the uterus and also to the alkaline mucosa of 
the digestive tract, but not to the acid slime of the vagina. His 
method of experimenting was as follows : A piece of the mucosa 
of the uterus was put on one side of a slide and a piece of the 
vagina on the other. A drop containing the sperm was placed in 
the middle of a cover-slip, and this put over the pieces on the 
slide. It was found that the sperm collected more on the side 
near the piece of the uterus, and from this Loew infers that they 
have been attracted to this side. In the light of the other experi- 
ments described above it will be clear, I think, that the greater 
accumulation of the sperm on one side by no means establishes the 
conclusion that they have been attracted to this side. Loew tried 
to show that filter paper saturated with alkaline substances acts che- 
motactically on the spermatozoa, in the sense that they move to- 
wards such substances, but, as in the preceding case, it does not nec- 
essarily follow because spermatozoa collect around or in certain 
substances, therefore they must have moved towards these sub- 
stances. The recent work of Jennings on the protozoans shows that 
their accumulations in certain areas is not due to the action of sub- 
stances that cause the individuals to swim towards those sub- 
stances, but on the contrary to their action being such that those 
individuals that enter areas containing these substances are unable 
to leave them. The result is the same as when the spermatozoa 
touch the jelly of the egg and stick to it, although the means by 
which the accumulations are formed in the two cases are entirely 
different. It would be interesting to see if spermatozoa may 
not behave towards certain solutions as do the protozoans. 

Self -Fertilization Induced by Artificial Means. 169 


It has been often assumed by embryologists that there exists 
some sort of attraction between the eggs and the spermatozoa of 
the same species. This idea would readily suggest itself to anyone 
who saw spermatozoa collecting in crowds around the eggs, but 
it by no means follows that this phenomenon is really due to an 
attracting substance emanating from the egg. The result may 
be due to the membrane of the egg, to which those sperma- 
tozoa stick that come accidentally into contact with it. In 
fact I have observed similar collections of spermatozoa in the 
ascidian around pieces of the body tissue, where the result had 
every appearance of being due to some sticky substance, exuding 
from the piece, rather than to an attraction exerted by the piece 
on the spermatozoa. 

Pfeffer's oft-qiioted experiment M'ith tlie antherozooids of ferns, 
liverworts, etc., appears to support the idea that the antherozo- 
oids are attracted to the malic acid that is present in the neck of 
the archegonia, but in the light of the recent experiments of Jen- 
nings and others, as to the way in which unicellular forms accu- 
mulate in a drop of acid, we can readily see that the results may 
have a very different interpretation from that usually given to 
them. Confining our discussion to the results obtained with the 
ascidians, I offer the following tentative analysis of the problem : 

The failure of the spermatozoon of Ciona to enter the egg of 
the same individual may be conceived as due to some physical ob- 
stacle. It Is conceivable that pores may exist In the egg-membrane, 
or even In the surface of the egg itself. This is the argument 
that Ptliiger^ used in the case of cross-fertilization of the frog's 
egg. If in the ascidian there existed a correlation of such a sort, 
that the size of a spermatozoon of a given individual Is always 
greater than the pores of the eggs of the same individual, then 
self-fertilization could not take place. That this Is not the real 
explanation Is shown by the fact that good spermatozoa are ap- 
parently capable of fertilizing the eggs of all other individuals. 
This would certainly not be the case if the exclusion of the sperma- 

1 Archiv. f. die gesammte Physiologic, XXIX., 1882. 

lyo T. H. Morgan. 

tozoon from the egg of the same individual was due to the size 
of the pores, because there would be eggs of some other individuals 
having pores as small or smaller. Another possibility that sug- 
gests itself is that the surface tension of the egg is of such a sort 
that it excludes the spermatozoa of the same individual, but this 
idea does not appear to give a satisfactory solution, for, aside 
from the fact that it is difficult to imagine how such a relation 
could exist, there would also occur cases in which the surface ten- 
sion of the eggs of other individuals would exclude certain sperm, 
and this does not appear to be the case. It is true that the ad- 
dition of the ether to the water may cause a difference in the 
surface tension of the egg, and it might be made to appear that 
this was the way in which the self-fertilization is effected in the 
ether-solutions, but I can not believe that this is the explanation 
of the results, because other experiments show that a considerable 
amount of ether is necessary to cause self-fertilization. 

It seemed to me that violent shaking might so affect the sur- 
face of the egg that self-fertilization might take place. A 
number of eggs from the oviduct were violently shaken for a 
few minutes in a small vial, and then sperm from the same in- 
dividual was added. No segmentation took place, and the pre- 
sumption is therefore that the eggs were not fertilized. 

Turning to the chemical side we find a number of possibilities 
that demand consideration. The inactivity of the immature sper- 
matozoa, and the lack of power of such sperm to fertilize the egg, 
their becoming active in certain solutions, and their power then 
to fertilize eggs that they did not fertilize before, as best shown 
in Cynthia, suggests that normally the eggs may secrete certain 
substances that make more active the spermatozoa, which then be- 
come capable of fertilizing the eggs. This view appears all the 
more attractive in the present case on account of the observed leth- 
argy of the spermatozoa of these ascidians, and the apparent con- 
nection in such cases between this condition and the impotence of 
such sperm in fertilization. Yet after careful consideration I am 
not prepared to advocate this view as the only solution, although I 
realize that it might be made to give the appearance of a ready 
explanation of my results. Not that this induced activity may not 

Self-Ferttlization Induced by Artificial Means. 171 

be one of the factors to be taken into consideration, but it is not, 
I think, the whole explanation. My reasons for regarding this 
view as insufficient are the following: It was found that sperm 
that appeared to be very little active was sometimes capable of 
cross-fertilizing the eggs of another individual. Possibly this 
may be due to somewhat greater activity induced by something 
secreted by the eggs of the other individual, yet on the whole I 
can not claim that direct obsei-vation gave any convincing evi- 
dence in favour of this assumption. More significant are the 
results of the experiment of mixing eggs from two Individuals, 
and subsequently fertilizing them with the sperm from one of. 
the Individuals. Half only of the eggs segmented, presumably 
those cross-fertilized. If some substance that makes the sperm 
active were really thrown out by the eggs, then we should expect 
that all the eggs would have been fertilized, unless indeed the se- 
cretion loses its power a short distance from the surface of the 
egg that secretes it; but this does not seem to be a probable Inter- 

A different point of view Is that the egg secretes some sub- 
stance that attracts the spermatozoa. On this view we must 
suppose that the substance secreted by the egg of Ciona has no 
attraction for the spermatozoa of the same individual. 

The little evidence that I have to' offer, based on experiments 
with ascidians. Is not favorable to this idea, that the cross-fertiliza- 
tion is due to some attractive substance secreted by the egg. In 
the species that I have examined there is no such marked 
accumulation of spermatozoa around the eggs as is seen in many 
other animals, and nothing in the behaviour of the cross- and self- 
fertilized egg to suggest that the difference in the results is due 
to an attraction in the one case, and to the absence of an attraction 
In the other. In other forms where there is a better opportunity 
for examining this question the most recent observations go to 
show, as has been pointed out in detail above, that there Is no suf- 
ficient evidence for the view that the egg attracts the spermato- 

Conversely, it may be supposed that the egg secretes some sub- 
stance that repels the spermatozoa of the same individual. I 

172 T. H. Morgan. 

observed nothing that would support such a conclusion, and this 
interpretation of the process would be foreign to what we find 
in general in connection with fertilization even in cases where the 
sperm of one species does not fertilize the eggs of another. 

We come now to a more subtile argument, and one that we 
are scarcely in position to discuss profitably in our present state 
of ignorance concerning the union of egg and spermatozoon. 
It may be assumed that there is some sort of "chemical affinity" 
between the egg and the spermatozoon that causes the two to 
unite when they come together. On this assumption we should 
have to suppose in Ciona that this affinity does not exist, or at 
least is less strong, between the egg and the spermatozoa of the 
same individual than between those of different individuals. Such 
a statement carries us no further, however, than the facts, and in 
the case of Cynthia we should have to assume that the affinity is 
so nicely balanced that sometimes the spermatozoon can unite, and 
sometimes it can not. In the case of Molgula the affinity must 
be assumed to suffice to bring about self-fertilization. Until we 
can give some more tangible form to this idea it does not appear 
to have any greater value, than the mere statement of the facts, 
and indeed may have less value, since it may give a wrong im- 
pression as to the real factors at work. 

Finally there might be advanced what may be called the electro- 
chemical hypothesis. The union of the egg and the spermatozoon 
may be supposed to be an electrical phenomenon, connected with 
a difference in the chemical composition of the two elements. The 
sperm head is almost pure nuclear chromatin, while the surface 
of the egg is protoplasmic. Possibly the spermatozoon and the 
egg have different electrical charges and unite with each other 
if brought near enough for the charges to become effective. But 
on this supposition it is not clear why the eggs and the sperm of 
the same individual would not unite. Here also we get no light 
on the absence of self-fertilization in Ciona. 

I have kept constantly in mind while at work on this problem 
the possibility that the spermatozoon may really enter the egg, but 
fail to develop there, or fail to start the development of the egg, 
because, coming from the same individual, it was not sufficiently 

Self -Fertilization Induced by Artificial Means. 173 

different In composition to supply the necessary stimulus. The 
ether might be supposed to make the sperm sufficiently different 
from the egg to start the cleavage, or the ether might itself supply 
the stimulus which is capable of starting the development of the 
egg after the spermatozoon has entered. 

The test of this view should be found in direct . observation 
of the eggs themselves. I prepared therefore a series of eggs 
of Ciona, some unfertilized for check series, others self-fertilized, 
but not put into ether, and others like the last, but put into ether. 

The difficulties of determining whether the spermatozoa can 
enter the eggs of the same individual, but fail to start the devel- 
opment, are greater than may appear at. first sight. The sperm 
head Is so minute that If after it entered no changes were af- 
fected In the protoplasm about it, Its presence might be readily 
overlooked, and since the spermatozoon of Ciona^ enters the egg 
in a granular zone that colors more deeply in certain stains than 
does the rest of the egg, the difficulty is thereby increased. Of 
course I have been on my guard against cases where the sur- 
rounding sperm have floated over the section, as sometimes hap- 
pens, or have been carried over it by a defect In the knife, and I 
have also been careful to exclude all cases where specks of foreign 
matter may have been on the slide, or in the fixative. There are 
also two further precautions to be taken. When the egg with- 
draws from the membrane and the test-cells are extruded, as it 
were, from the outer zone of the egg, the protoplasm is some- 
times drawn out in mamiliform processes that stain deeply and 
resemble the entrance cone formed by the spermatozoon pene- 
trating certain eggs. Even when the protoplasm does not pro- 
trude, deeply staining spots are generally present and are espe- 
cially obvious after Iron haematoxylin. Careful staining with 
Delafield's haematoxylin shows clearly that these spots have noth- 
ing to do with the entrance of spermatozoa. Furthermore these 
spots are found in unfertilized eggs. After iron haematoxylin 
minute deeply staining bodies, flattened against the outer surface 
of the egg, can generally be found, and these strongly suggest 
spermatozoa. That they are not such is shown by their presence 
in unfertilized eggs, and by their absence after the Delafield 

174 T. H. Morgan. 

stain. I mention these points because they might easily lead one 
after only a casual examination to conclude that spermatozoa 
enter the eggs. My best results have been obtained by drawing 
out the iron haematoxylin until the protoplasm has lost all of its 
color, or better still by using the Delafield stain, and also thor- 
oughly extracting the color from the protoplasm. 

Although I have examined a large number of preparations I 
have not seen a single definite case without ether in which a 
spermatozoon has entered the egg of the same individual. Diffi- 
cult as it admittedly is to be absolutely certain on this point, yet if 
the spermatozoa had entered and had begun to enlarge I feel 
certain that I should have detected their presence. That un- 
developed sperm-heads may be present I must admit as a pos- 
sibility, but I have not detected them, and believe that I should 
have been able to do so were they present. It is also a point of 
some importance that I have not found any spermatozoa within 
the egg membrane, although quantities of them may lie outside. 

There is a further point in this connection, the importance of 
which I did not appreciate until I had closed the experimental 
part of my work. In the eggs of many animals a change takes 
place In the egg, after the penetration of one spermatozoon, of 
such a sort that the entrance of more spermatozoa is prevented, 
I have found In Ciona that, after the sperm has stood with the 
eggs of the same individual and has failed to fertilize them, these 
eggs could still be readily fertilized by spermatozoa from another 
individual. If a spermatozoon of the same individual really 
enters the egg it does not In consequence bring about such a change 
in the egg that other spermatozoa can not enter, and therefore 
many spermatozoa of the same Individual from which the eggs 
were taken should be expected to gain entrance, but I am quite 
certain that this, at least, does not occur. From this consideration 
also It may be Inferred that the spermatozoa do not normally pene- 
trate the eggs of the same individual. 

In the light of these observations it seems probable that whenever 
a spermatozoon enters the egg, the egg begins to develop regard- 
less of whether the spermatozoon comes from the same or from 
another Individual. The ether must therefore induce a change 

Self -Fertilization Induced by Artificial Means. 175 

of some sort that directly effects the entrance of the spermatozoon 
into the egg, and at present I see no other interpretation that is 
left than that this entrance is due to the greater activity of the 
spermatozoon that causes it to overcome some resistance, either 
on the surface of the egg itself, or in the membrane surrounding 
it. The nature of this resistance I did not detect, and this must 
be the next step in the analysis. One method by which this view 
may be tested is obvious, and has already been referred to. The 
spermatozoa made active by sea-water must be placed in an ex- 
tract of the eggs (or body-tissues) of the same individual, and 
then, after a time, the eggs of another individual added. On 
the hypothesis these eggs should be less likely to become fertihzed 
than eggs placed directly in contact with the fresh sperm. 

It has been found that certain substances secreted by the glands 
of the reproductive organs of the male mammal arouse the sper- 
matozoa to greater activity. It has also been found that many 
other substances have a similar effect on spermatozoa. It would 
be equally interesting to discover if the secretions of other parts 
of the genital ducts of the male or of the receptacula of the fe- 
male, when such are present, may not bring the spermatozoa 
to rest, or keep them quiescent until some other exciting agent 
arouses them. It seems almost certain that this must be the case 
in those animals in which the spermatozoa of the male are stored 
up in receptacula of the female, as for instance in the honey bee, 
or in such a hermaphroditic animal as the earthworm. The 
length of life of the spermatozoa in some of these forms would 
seem to make some assumption of this sort necessary. Experi- 
ments can easily be made that would decide this question. Kol- 
liker has shown, in fact, that water quiets the spermatozoa of 
mammals without killing them. 

In the ascidians it is probable that the spermatozoa in the vas 
deferens are quiescent. It is significant that in these hermaphro- 
ditic forms the oviduct in which the eggs are stored takes a course 
parallel to the male duct. Possibly the proximity of the two 
ducts may be connected with the lack of power of self-fertilization 
of the eggs, because the egg may be saturated with the same 
substances that keep the sperm quiescent. It may be, however, 

176 T. H. Morgan. 

that this relation is more fundamental, and the particular substance 
is one peculiar to the whole body. That the reaction must be 
something quite specific is shown by the fact that the spermatozoa 
are able to enter eggs of any other individual. 

It appears probable that of all the different substances that 
excite the spermatozoa to activity the secretions of the glands 
connected with the male reproductive organs may be the most 
efficient. From a statement of KoUiker's it seems not improbable 
that the substance secreted in the glands of one species may be 
also efficient for the spermatozoa of other species. Whether 
by the use of the substances from the glands of another mammal 
it might not be possible to excite human spermatozoa to greater 
activity and thus assist materially in bringing about fertilization 
in cases where the impotence is on the side of the male remains 
to be examined. There is here a question that may have an im- 
portant practical aspect. 

The lack of power to self-fertilize in plants may also be due 
to the inability of the pollen tube to penetrate sufficiently far into 
the stigma and style. It appears that penetration does actually 
begin in some cases that have been observed, but possibly the 
growth may be arrested further down in the style. The pre- 
potency of other pollen would then find its explanation in the 
more rapid growth of this foreign pollen. Here again is an op- 
portunity for future work.^ 

In attempting to formulate a theory to account for the deter- 
mination of sex, Castle assumes that there are two kinds of sperma- 
tozoa, male and female, and that there are also two kinds of eggs, 
male and female. He also assumes that a female egg can be 
fertilized only by a male spermatozoon and that a male egg only 
by a female spermatozoon. I have already pointed out elsewhere^ 
that my results do not support this assumption. Castle appealed 
to the case of Ciona as one in favour of his contention, for the 
eggs here can not be fertilized by the sperm of the same individual. 
It is not explicitly stated to the contrary, and the reader might be 
led to infer from the context that in Ciona all the eggs and all 

1 The experimmts of Myoshi should be especially considered. 

2 Popular Science Monthly. Dec, 1903. 

Self -Fertilization Induced by Artificial Means. 177 

the spermatozoa of one individual must be supposed to be male, 
and in another individual the reverse; but certainly this is not the 
case, and could not have been Castle's meaning, for if it were so 
then half of the individuals would be infertile with the sperm of 
the other half, and this is not so. I have pointed out that 
my results with ether, etc., do not support Castle's assumption, 
although it might, of course, be claimed that the ether causes 
the spermatozoa to lose, as it were, their homosexual repugnance. 
However this may be, I have found that no such lack of power 
to self-fertilize is found in some other ascidians, as in Molgula 
for example. If my supposition is correct, that self-fertilization 
in Ciona is due to the presence in the eggs of some substance that 
brings the spermatozoa to rest, the whole question assumes a very 
different aspect and does not appear to have any connection with 
the question of the determination of sex. 


BoRDET. — Contribution a I'Etiide de I'lrritabilite des Spermatozoides chez les 

Fuccacees. Bull, de I'Acad. Belgique 3W(?. ser. XXVII. 1894. 
BuLLER, A. H. — The Fertilization Process in Echinoidea. Report 70. Meeting 

Brit. Assoc, ipoi. 

Is Chemotaxis a Factor in the Fertilization of the Eggs of Animals. 
Quart. Journ. Micro. Sc, XLVI. 1903. 
BuLLER, A. H. R. — Contributions to our Knowledge of the Physiology of the 

Spermatozoa of Ferns. Ann. of Botany, XIV. 1900. 
Castle, W. E.— The Heredity of Sex. Bull. Mus. Comp. Zool.. XL. 1903. 
Dewitz, J. — tjber Gesetzmassigkeit in der Ortsveranderimg der Spermatozoen 

und in der Vereinigung derselben mit dem Ei. Archiv. f. die gesammte 

Physiologic. XXXVIII. 1886. 
V. DuNGERN, E. — Die Ursachen der specifitat bei der Befruchtung. Centralbl. f. 

Phys. XIIL 1901. 

Neue Versuche zur Physiologic der Befruchtung. Zeitschr. f. allgem. 
Physiologic. I. 1902. 
IvANOFF. — Journal de Physiologic et de Pathologic general. II. 1900. 
KoLLiKER, A. — Physiologische Studien ucbcr die Samcnflussigkeit. Zeitschr. f. 

wiss. Zool. VII. 1856. 
Kraft, H. — Zur Physiologic des Flimmerepithels bei Wirbelthieren. Archiv. f. d. 

gesammte Physiologic. XLVII. 1890. 
LiDFORSS. — iJber den Chemotropismus der PoUenschlauche. Ber. d. Deutsch. 

Bot. Gcsell. XVII. 1895. 

178 T. H. Morgan. 

LoEW, O. — Die Chemotaxis der Spermatozoen im weiblichen Genitaltract. Sitz- 
ungsber. d. Wiener Akad. ; Math.-naturw. CI. CXI. 1903. 

Massart, J. — Sur rirritabilite des Spermatozo'ides de la Grenouille. Bull, de 
I'Acad.roy. de Belgique. 2'me Sir. XV. 1888. 

Sur la Penetration des Spermatozoides dans I'Oeuf de la Grenouille. 
Bull, de I'Acad. roy. de Belgique. Zme. Ser. XVIII. 1889. 

MiYOSHi, C. — tjber Reizbewegungen der Pollenschlauche. Flora. LXXVIII. 1894. 

MoLiscH, H. — tJber die Ursachen der Wachtumsrichtunger bei Pollenschlau- 
chen. Sitzungsber. d. k. Acad. d. Wiss. in Wien. 1889, 1893. 

Morgan, T. H. — Recent Theories in Regard to the Determination of Sex. Pop- 
ular Science Monthly. Dec, 1903. 

Pfeffer, W. — Locornotorische Richtungsbewegungen durch chemische Reize. 
Untersuchungen aus d. Bot. Inst, zu Tiibingen. I. 1884. 

Steinach, E. — Untersuchungen zur vergleichenden Physiologic der mannlichen 
Geschlechtsorgane insbesondere der accessorischen Geschlechtsdriisen. 
Archiv f. d. gesammte Physiologic. LVI. 1894. 

Strasburger, a. — Das botan Prakticum, 2 Auf 1887. Page 402. 





From the Rudolph Spreckels Physiological Laboratory of the 
University of California. 

In previous publications^ it was shown that subcutaneous or 
intravenous injections of small quantities of solutions of certain 
salts, including the saline purgatives, produce not only increased 
peristalsis, but also an increased secretion of fluid into the in- 
testine. This was found to be true also when the solutions were 
applied locally to the peritoneal surfaces of the intestine. It was 
suggested that the main actions of saline purgatives consist in 
the production of increased peristaltic movements, and of in- 
creased secretion of fluid into the intestine; and that the semi- 
fluid foeces which are produced by saline purgatives are the 
result not of decreased power of absorption by the intestine, 
but of an increased secretion of fluid into the intestine. It was 
further shown that the administration of calcium or magnesium 
chloride tends to suppress the peristaltic movements and the secre- 
tory activity of the intestine. Attention was specially called to the 
marked action of barium chloride in the production of violent peri- 
staltic movements and ringlike constrictions in the intestine, and 
also in the production of an increased flow of fluid into the intes- 
tine. It was also pointed out that the production of these activi- 
ties in the intestine by the purgative salts, and their suppression 
by calcium and magnesium is analogous to the production and 
suppression of rhythmical contractions in voluntary muscles de- 

1 A preliminary report of these experiments was published in the University 
of California Publications, Physiology, January 15, 1904, Vol. I., No. 10, p. 81. 

2 MacCallum, J. B. — American Journal of Physiology, Vol. X., No. III., 
p. loi, 1903, and Vol. X, No. V, p. 259, 1904. 

i8o Joliii Bruce MacCallinn, M. D. 

scribed by Loeb^. The antagonism which has been shown by 
Loeb to exist between the actions of many sodium salts on the 
one hand, and calcium and magnesium salts on the other was 
further illustrated by these experiments. 

It seemed possible then in the light of these facts that the activ- 
ities of the kidney might be controlled in the same way as those 
of the intestine. Since it is well known that many sodium salts 
have a distinct diuretic action, it seemed conceivable that calcium 
or magnesium might act as an antidiuretic. In order to decide 
this point I have made a series of experiments in which I have 
found that the relation of many of the salts to the activity of the 
kidneys is similar to that which they bear to the glandular activity 
of the intestine. 


The following experiments were carried out mainly on rabbits; 
a few dogs also were used. In all cases morphine was given 
as an anaesthetic. The rabbits received 3-5 cc. 1% solution of 
morphine hydrochlorate subcutaneously ; the dogs in addition 
to this dose of morphine were given ether when necessary. 

The urine was collected by catheterising the ureters or by tying 
a cannula in the bladder. The latter method was employed in 
all cases except those in which it was necessary to observe the dif- 
ference between the amounts secreted by the two kidneys. In 
placing a cannula in the bladder a small incision was made in the 
abdominal wall. The bladder, which usually contains a consid- 
erable quantity of urine, was then lifted out of the body cavity, 
and the abdominal wall sewed up around the neck of the bladder 
so that the intestines could not be forced out. A purse-string su- 
ture was then made in the fundus of the bladder and an incision 
made in the bladder wall within the suture. In this way the urine 
could be removed, and the cannula securely tied in. Care was 
taken to allow no urine to collect in the bladder, so that the meas- 
urements given in the tables represent all the urine that was se- 
creted during each period. The simple catheterisation of the 
bladder through the urethra may be easily performed in rabbits, 

1 Loeb, J. — Festschrift fiir Fick, 1899; Archiv. fiir die gesammte Physiologic. 
1902, XCI, p. 248. 

Influence of Calcium and Barium on the Kidney. i8i 

but this method Is unsatisfactory when it is necessary to obtain 
the exact amount of urine secreted in a given time since it is im- 
possible to tell whether the bladder is at any time entirely empty. 
Solutions of the salts whose actions were tested were introduced 
into the body intravenously. In rabbits a hypodermic needle was 
placed in a vein of the ear; in dogs the fluid was forced into one 
of the superficial veins of the lower limb. When small quantities 
were injected a hypodermic syringe was used; when larger amounts 
were introduced a pressure apparatus was employed. This was 
the apparatus commonly used in injection work, consisting of a 
pressure bottle connected on one side with a water tap, and on 
the other with a graduated bottle containing the solution to be 
injected. In this way a constant pressure could be obtained, and 
the quantity of fluid injected in a unit of time accurately con- 
trolled. By causing the fluid to pass through a coil of rubber 
tubing immersed in hot water before reaching the needle, the 
solution could be kept constantly at the body temperature. Some 
of the details of the apparatus were suggested to me by an ap- 
paratus used by Dr. M. H. Fischer in this laboratory. For such 
Infusions into the blood only 'Vs solutions were employed; In 
subcutaneous injections stronger solutions were used. Except In 
those cases where It was necessary to obtain the eftect of the salt 
on the normal flow of urine, the secretion was considerably raised 
and kept constant by the uniform infusion of ""/g NaCl solution 
throughout the experiment. The effect of the other salts was 
then obtained by allowing small quantities of '"/s solutions to flow 
into the vein along with the NaCl solution. In other Instances 
these salts were Injected into a vein of one ear while the NaCl solu- 
tion was at the same time flowing Into the opposite ear. In these 
experiments the ear of the rabbit was securely tacked to the board, 
and the needle kept from slipping out of the vein by means of 
bull-dog forceps. 


The results of the experiments on the actions of calcium and 
barium may be best seen In the following tables : 

I. Dog — small terrier — cannula placed In right ureter. 

Urine secreted In ist lo minutes 3.6 cc. 

2d 10 minutes 3-6 " 

1 82 John Bruce MacCallum, M. D. 

8 cc ""/s CaClg injected into vein of leg. 

Urine secreted in ist lo minutes 2.4 cc 

2d 10 minutes 2.2 " 

3d 10 minutes 1.8 " 

4th 10 mniutes 1.6 " 

5th 10 minutes i-4 " 

10 cc ""/s sodium citrate injected subcutaneously. 

Urine secreted in ist 10 minutes 1.6 " 

2d 10 minutes 2.3 " 

3d 10 mniutes 3.1 " 

4th 10 minutes 3.6 " 

In this case the secretion of urine gradually decreased after the 
injection of calcium chloride until the amount collected in a unit 
of time was less than half of the initial amount. The addition of 
sodium citrate to the blood counteracted this effect so that the 
rate of secretion again approached the normal. These effects 
are more striking when the quantity of urine secreted is increased 
by the introduction of normal salt solution into the blood as shown 
in the following experiment: 

2. Rabbit — cannula placed in bladder. No urine flowed in 
the first or second periods of 10 minutes before the NaCl solution 
was injected. 

Salts other than "Vg NaCl in- Urine in 
Time. NaCl injected. jected in cc. cc. 

10.10 10 

10.15 10 

10.20 5 0.5 

10.40 10 0.8 

11.00 10 0.5 

11.20 5 i.o 

11.40 10 2.8 

12.00 10 6.0 

12.00 5 cc Vs CaCla intravenously 

12.05 5 cc ^"/s CaClg subcutaneously 

12.20 5 0.2 

12.40 10 1.8 

i.oo 10 0.8 

1 .00 5 cc ^/g sodium citrate intravenously 

1.20 10 2.2 

1.40 5 3.6 

Influence of Calcium and Barium on the Kidney. 183 

In this experiment, although the flow of urine has been consider- 
ably increased by the injection of "Vs NaCl, the introduction of 
CaClo markedly suppresses the secretion. The flow of urine remains 
small for an hour, although a somewhat greater quantity of fluid 
is forced into the blood than in the previous hour. This suppres- 
sion of urine is at once counteracted by the injection of sodium 

The following table (3) which represents only the latter half 
of an experiment shows roughly the duration of the action of 
smaller doses of calcium. 

3. Rabbit — cannula in bladder — injections intravenous. 

Salts other than Vg NaCl in- Urine in 
Time. NaCl injected. jected in cc. cc. 


1.40 150 64.5 

1.45 10 6.G 

1.50 10 5.6 

1.55 10 6.2 

2.00 10 7.4 

2.05 10 9.5 

2.05 5 cc Vs CaCls 

2.10 5 2.2 

2.15 10 0.8 

2.20 10 1.2 

2.25 10 1.6 

2.30 10 2.8 

2.35 8 3.0 

2.40 5 4.5 

2.45 o 4.8 

2.50 o 5.1 

2.55 O 6.2 

As shown here and in other experiments, the action of calcium 
is only temporary. I have found also that magnesium chloride 
in many cases has cin antidiuretic action similar to that of calcium 
chloride. The suppression of urine, however, is not so marked 
as with calcium. 

As shown in the following experiment (4) barium chloride in 
very small doses has a strong diuretic action. Although it is 
much more powerful in this respect than sodium citrate, the in- 
creased flow of urine which it causes may be suppressed by the 
injection of calcium chloride. 

184 John Bruce MacCallum, M. D. 

4. Rabbit; injections intravenous. 

Salts other than ""/g NaCl in- Urine in 

Time. NaCl injected. jected in cc. cc. 


10.30 20 

10.40 10 

10.50 20 1.2 

11.00 -: 32 2.8 

II. 10 28 5.8 

1 1.20 20 6.1 

11.30 10 8.2 

11.40 10 8.3 

11.40 yi ccVs BaCls 

11.50 10 14.4 

12.00 10 18.0 

12.10 10 12.4 

12.10 >^ cc '"/s BaCls 

12.20 10 18.4 

12.30 10 16.4 

12.30 5 cc ""/s CaCL 

12.40 10 8.6 

12.50 10 4.0 

1. 00 10 2.0 

1. 10 ID 2.4 

1.20 5 3.4 

1.20 J4 cc Vs BaCla 

1.30 8 6.4 

1.40 10 8.2 

1.50 10 8.6 

1. 50 ^ cc ™/8 BaCla 

31.55 10 i.8( 

1 2.00 10 0.6 f 

2.10 10 0.0 

2.20 0.0 

2.30 0.0 

In the uniform injection of considerable quantities of normal 

salt solution into the blood, the flow of urine, after about an 
hour, becomes fairly constant. If an average amount of i cc. 
in I minute be introduced, the secretion of urine during the first 

two or three hours is usually slightly less than the amount of 

fluid injected. After this time, when no other salts have been 
added, the quantity injected and the quantity secreted may become 

Influence of Calcium and Barium on the Kidney. 185 

approximately equal. As shown in experiment 4 however the addi- 
tion of a minute quantity of BaCL (less than y^ cc ""/§ solution) to 
the blood causes the flow of urine to increase markedly, so that the 
quantity secreted in a unit of time is far in excess of the quantity 
of fluid introduced into the blood. If, however, while this active 
secretion is going on, 5 cc. '"/g CaClg solution be injected into 
the blood, the flow of urine rapidly decreases, although the total 
quantity of fluid added to the blood remains constant. The fur- 
ther addition of BaCU again increases the secretory activity so 
that the quantity secreted in 10 minute periods which has fallen 
from 16.4 to 2, under the influence of CaCL is again raised to 8.6 
by the injection of the barium salt. An apparently contradictory 
thing, however, happens when a larger amount of barium chloride 
is suddenly added to the-blood. As shown in the foregoing table, 
while yi cc. BaCL largely increases the urinary secretion, the 
injection of ^ cc. in addition to that already present, causes an 
entire cessation of the flow of urine. In some cases this suppres- 
sion of the flow of urine is quite abrupt; in other instances it is 
more gradual, a few drops of urine flowing from the cannula at 
intervals. As shown in the following experiment, the injection 
of CaCL sometimes counteracts this action of larger doses of 
BaClg and causes the urine to flow again. 

5. Rabbit — cannula in bladder; injections intravenous. 

Salts other than "Vg NaCl in- Urine in 

Time. NaCl inj"ected. jected in cc. cc. 


10.00 10 

10.15 15 

10.30 15 

10.45 15 

11.00 15 

11.00 I cc Vs BaCU + 4 cc Vs NaCl 

II. 15 10 

11.30 15 

11.45 15 

11.45 5 cc. Vs CaCla 

12.00 10 

12.15 • .15 

12.30 15 








. I 



2 , 











1 86 John Bruce MacCallum, M. D. 

In this case i cc "/s BaClg gradually suppresses the flow of 
urine, and no trace of the strong diuretic action of barium is seen. 
And, further, calcium chloride has here an action which seems at 
first glance entirely opposed to that which it ordinarily has. As 
shown in the previous experiments, calcium characteristically sup- 
presses the secretion of urine. In this case the flow of urine 
increases after its administration. These apparent contradictions 
may be explained in the following way. In discussing the actions 
of calcium and barium on the intestine, it was pointed out that 
barium chloride, like the other saline purgatives, affects the in- 
testine in two ways, namely, by increasing the peristaltic move- 
ments and by increasing the secretion of fluid into the lumen. At- 
tention was further called to the violent character of the muscular 
contractions in the intestine caused by barium, which may so con- 
strict the lumen of the intestine that fluid cannot pass from one 
part to another. It was also shown that calcium to some extent 
counteracts the action of barium both on the muscle, and on the 
glands of the intestine. It seems therefore probable that the 
increase in the flow of urine caused by small doses of barium 
chloride {yi cc. ""/« solution) is due to an increase in the secretory 
activity of the kidney entirely analogous to that which is pro- 
duced in the intestine by the same salt. The cessation of the flow 
of urine however which follows the administration of larger doses 
of barium chloride (icc. "/g solution) is in all probability due 
to the action of the barium on the muscle coats of the urinary pas- 
sages, especially those of the calyces and pelvis of the kidney, and 
those of the ureter. Since all of these various parts of the urinary 
passages are surrounded by thick, circular and longitudinal muscle 
coats, not unlike those of the intestine, it seems conceivable that 
a strong contraction of these coats, such as barium is capable of 
causing in the intestine might effectually shut off the lumen so 
that no urine could pass. Furthermore the action of calcium in 
renewing the flow of urine under these circumstances is quite 
analogous to its action in suppressing the peristaltic movements 
or in relieving the constrictions in the intestine caused by barium. 
The actions of calcium and barium which are shown in Table 5, 
are on the muscle coats of the urinary passages. It is quite conceiv- 

Influence of Calcium and Barium on the Kidney. 187 

able however that this action of calcium may coexist with its char- 
asteristic action in diminishing the secretory activity of the kidney. 
In both the intestine and the urinary apparatus (kidney, and 
urinary passages) barium stimulates the glandular and the mus- 
cular tissues to activity. Calcium on the other hand uniformly 
suppresses these activities. 

It must be pointed out however that the suppression of the 
flow of urine which follows a relatively large dose of barium 
chloride cannot always be relieved by calcium. As was found to 
be true In the intestine, the action of barium Is seldom completely 
counteracted by calcium. In many cases the barium stops the 
flow of urine entirely so that it is not possible to start it again. 
This is shown In the following experiment (6) where relatively 
large quantities of calcium chloride are incapable of reestablishing 
the flow of urine. This naturally suggests the idea that the large 
doses of barium may stop the secretion of urine by injuring the 
cells of the kidney, or perhaps Indirectly by a constricting influ- 
ence on the blood vessels. These possibilities must be taken into 
consideration; but the fact that calcium sometimes causes the 
urine to flow again after it has been Inhibited by barium speaks 
strongly In favor of the theory advanced above, that the inhibiting 
action of barium on the flow of urine Is an action on the mus- 
cular tissue of the urinary passages. 

6. Rabbit — cannula In bladder; Injections Intravenous. 

Salts other than '"/s NaCl In- Urine in 

Time. NaCl Injected. jected In cc. cc. 


10.20 23 

10.30 20 

10.40 25 

10.50 28 

11.00 20 

II. 10 16 

11.20 10 

11.30 10 , 

11.40 12 

11.50 15 

1 1.5 I ^ ccVs BaCl2 

11.55 i-o 



I , 


2 , 











. 2 

1 88 John Bruce MacCaUum, M. D. 

12,00 15 ,. 14-4 

12.00 . Yz cc 

12.05 — 

12.10 — 

12.20 — 

12.32 5 cc "Vj 

/s oaCla 







; CaCl^ 










12.50 5 cc 

1. 00 — 

1. 10 — 

1.20 - 

It will be noticed in this experiment (6) that immediately after 
the injection of ^ cc. ""/g BaClg solution there is a marked diminu- 
tion in the flow of urine followed within a few minutes by a very 
considerable increase. This partial cessation of the flow imme- 
diately following the injection is due probably to a temporary 
action of the barium on the muscle coats of the urinary passages. 
The subsequent increase is the result of the diuretic action of 
barium on the kidney as described above. 

In considering the actions of calcium and barium we must 
therefore take into account not only their influence on the glandular 
tissue, but also their effect on the muscular tissue of the body. 
In all cases these salts are antagonistic in their action ; and their 
influence on the secretory activity of the kidney and on the flow 
of urine is entirely analogous to their influence on the glandular 
and muscular activities of the intestine. With regard to its action 
on the kidney calcium chloride may be properly termed an anti- 

Attention must be again called to the extremely poisonous nature 
of barium chloride. A subcutaneous injection of 3 cc 'Vs BaCU 
solution is usually sufficient to kill a rabbit. Intravenously it 
should always be injected with four or five times its volume of 
'Vs NaCl solution. 

I. In dogs and rabbits the quantity of urine secreted in a unit 
of time may for a time be markedly diminished and in some cases 
almost entirely inhibited by the Introduction of calcium chloride 
into the circulation. 

Influence of Calcium and Barium on the Kidney. 189 

2. Calcium chloride diminishes not only the normal flow of 
urine, but also that which is caused by the administration of saline 
diuretics. For example, the rate of secretion which has been 
largely increased by the intravenous injection of normal salt so- 
lution may be temporarily lessened to a marked extent by the in- 
troduction of CaClo into the blood. 

3. In all cases "/g solutions were used, and ""/g NaCl solution 
was introduced into the blood at a constant rate throughout the 
experiments. After a short time the rate of secretion became 
constant. It was then found in rabbits, that the addition of a 
small quantity of BaCL (^ cc ""/g solution) to the blood causes 
a marked increase in the flow of urine, so that the amount of 
fluid secreted may considerably exceed that which is introduced 
mto the blood during the same period of time. 

4. This action of barium is counteracted by the injection of 

5. If a larger quantity of BaCL ( i cc ""/g solution) be added 
to the blood, the flow of urine ceases and often complete anuria 
ensues. In some cases the injection of CaCL abolishes this in- 
hibitory action so that the urine flows again. Usually however 
the action of barium persists. 

6. The fact that barium when given in smaller and in larger 
doses may thus apparently have opposite effects on the flow of 
urine may be explained by analogy with Its action on the intes- 
tine. Barium chloride causes not only an increase in the secretion of 
fluid into the intestine, but also active peristaltic movements, and 
violent local constrictions of the Intestine. Similarly very small 
doses of BaClg increase the secretory activity of the kidney. It 
seems^ probable however that the cessation of the flow of urine 
which follows the injection of larger quantities of the salt is due 
not to an inhibition of secretion, but to the action of the barium 
on the muscular coats of the urinary passages, especially those 
of the calyces and pelvis of the kidney and those of the ureter. 
This action would bring about a constriction of the tubes and a 
closure of the lumen. The fact that calcium counteracts both 
effects of the barium supports this explanation. 

190 John Bruce MacCallum, M. D. 

7. The influence of calcium and barium on the flow of urine 
is in every way analogous to their action on the intestine, which 
I have previously described. The suppression of the urinary 
secretion by calcium is also analogous to the suppression of twitch- 
ings in voluntary muscles by calcium, which has been described 
by Loeb. 

In conclusion it is a pleasure to thank Professor Loeb for the 
interest which he has taken in these experiments. I am indebted 
also to Dr. Theo. C. Burnett, who has assisted me in many of 
the experiments. 






Associate Professor of Anatomy, the Johns Hopkins University, Baltimore. 


F. H. BAETJER, M. D., 
Director of the Roentgen Apparatus, the Johns Hopkins Hospital, Baltimore. 

The great capacity of regeneration possessed by fresh-water 
planarians is well known because of the number of investigations 
lately devoted to the subject.^ Complicated portions of the body, 
like the head and pharynx, when removed, are restored within a 
few days or weeks. In several species a new individual may 
be regenerated from a very small, isolated piece. We have found 
that this power of regeneration may be completely destroyed by 
exposing planarians to the action of the Roentgen rays. 

Our experim.ents have been conducted upon two species of 
planarians which have especial regenerative capacity, P. maculata 
and P. lugubris.. Specimens were placed in shallow, open dishes 
about ten to fifteen centimeters below the vacuum tube and were 
exposed from ten to twenty minutes each day for varying periods 
of time. We made use of a twenty-centimeter coil with an inter- 
rupter of the electrolytic type, a ten-Inch coil with a mechanical 
Interrupter, and several styles of vacuum tubes. The vacuum of 
each tube was so arranged that rays of "medium-soft" quality 
were obtained. 

1 In addition to the literature quoted in Morgan's "Regeneration," New York, 
1901, articles have appeared as follows: F. R. Lillie, American Journal of 
Physiology, VL, p. 129, 1901 ; E. Schultz, Zeitschrift f. wissenschaftliche Zo- 
ologie, LXXII, p. I, 1902; N. M. Stevens, Archiv. f. Entwichelungsmechanilk 
XIIL, p. 396, 1901 ; T. H. Morgan, Archiv. f. Entwichelungsmechanik, XIIL, p. 
179, 1901 ; Biological Bulletin IIL, p. 132, 1902 ; H. F. Thacher, American Nat- 
uralist, XXXVL, p. 633, 1902; C. R. Bardeen, Biological Bulletin III., 262, 1902; 
Archiv. f. Entwichelungsmechanik, XVP., p. i, 1903. 

192 Charles Russell Bardeen, M. D. 

Our first experiments were upon worms from each of which 
the anterior region of the body was removed Immediately before 
the first exposure. The cut edges became closed In by muscular 
contraction and the extension of epithelium In the usual manner. 
For some days there was a slight production of new tissue near 
the cut surfaces, but this soon ceased. No new heads were pro- 
duced and no new pharynges. In one specimen of P. maculata, 
however, an Imperfect eye was regenerated on the left side at the 
junction of the old tissue with the new, and a very small eye-spot 
appeared on the right side, but the anterior end of the piece 
at no time assumed the normal shape of a head. The specimens 
were subjected to thirteen exposures. All died between the twen- 
tieth and twenty-second days after the first exposure. Control 
specimens regenerated In the usual manner. A "tail-piece" of 
P. maculata had become a worm of perfect form and proportions 
on the fifteenth day after the operation, and a "tail-piece" of 
P. lugiibris had regenerated a well proportioned new head and a 
new pharynx at that time. 

Several experiments were made to test the effect of the 
rays on uninjured worms. Specimens were exposed from 
twelve to eighteen times and were then kept under as hy- 
gienic conditions as possible. Some of these Individuals lived 
a month after the first exposure. During this period they 
reacted normally to light, to mechanical and to chemical (food) 
stimuli. Microscopical preparations made from a few specimens 
at varying periods after the last exposure showed no marked 
alterations In the muscular, nervous and Intestinal apparatus. The 
cutaneous epithelium seemed to be normal except for a few areas 
where the cells were shorter and broader than usual. The cUIa 
of the ciliated cells were Intact. In specimens with well-developed 
genitalia the cells of the testes showed no karyoklnesls. On the 
contrary most of them seemed to be undergoing a degenerative 
change. In corresponding control specimens mitosis was most 
active In these cells. No clear Instances of nuclear degeneration 
were observed. Death In these specimens resulted from a de- 
generative process which began In the region of the head and 
extended slowly back. This degeneration was probably parasytic 

Inhibitive Action of Roentgen Rays. 193 

in nature and seemed due possibly to insufficient protection offered 
by the thinned epithelium. Young individuals, of which we had 
several hatched out from an egg-capsule of P. lugubris, a few 
weeks before the experiments began, were affected like the ma- 
ture specimens, but more quickly. 

From worms exposed from ten to fifteen times to the rays 
pieces were cut immediately in some instances, and in some in- 
stances a week or ten days after the last exposure. In each instance 
the cut surfaces were closed by muscular contraction and mechan- 
ical extension of the surface epithelium, but in no instance was 
there subsequently seen any sign of the production of new tissue 
at the cut surface or in a region of the body where under normal 
conditions a new pharynx would be formed. Microscopical sec- 
tions o'f a specimen killed within the second twenty-four hours 
after the isolating cut was made showed no signs whatever of 
cell-division, either direct or indirect. In control specimens mitosis 
was most active in the tissue-forming parenchymal cells at this 
period. In the exposed specimens the epithelium where it had 
extended out to cover a cut surface remained a flat, thin mem- 
brane as long as the specimen lived. In the control specimens 
it was quickly restored to its normal columnar form. The ex- 
posed Individuals lived for from twenty-five to thirty days from 
the time of the first exposure. One piece of P. maculata, from 
which the head had been remov^ed, lived for forty-one days. Re- 
actions to light and to mechanical and chemical stimuli seemed 
normal in all the specimens. 

From these experiments it is evident that the Roentgen rays 
have a powerful inhibitive effect upon cell-reproduction in pla- 
narians. It may be entirely stopped by sufficient exposure. No 
effect was noticed in the physiological activities or in histological 
structure of the highly differentiated tissues such as those of the 
nervous system and the musculature. The effects of the rays do 
not appear for some days after the first exposure. Thus there 
is a slight production of regenerative material at a cut surface 
in a specimen sectioned before exposure to the rays. The sub- 
sequent differentiation of an imperfect eye in one specimen in- 
dicates that the rays have effect not so much upon tissue differ- 

194 Charles Russell Bardecn, M. D. 

entiation as upon cell-reproduction. The spreading out of the 
surface epithelium so as to cover a cut surface, whereby columnar 
epithelium becomes transformed into pavement epithelium also 
indicates this. Death in exposed specimens may possibly be due 
to a necessity on the part of the organism for a certain amount 
of cell-reproduction. 

The effects of the Roentgen rays on planarians thus tend to 
support the view of those investigators who regard its effects 
upon the tissues of other animals as due primarily to its action 
on cells capable of reproductive activity. Scholtz in his excellent 
clinical and experimental studies on the effects of X rays on the 
mammalian skin^ concludes that both the nuclei and the cell- 
protoplasm of the epithelial cells are injured by the rays, but 
that the effect on the connective tissues, elastic tissue, muscu- 
lature and cartilage is very slight if any. The skin on both sides 
of a rabbit's ear may be affected when it is exposed to rays on one 
side only. 

The effect is not, however, a direct one upon the actual process 
of cell-division. This is shown in planarians by the production 
of tissue at a cut surface during the first few days of exposure 
to the rays. It is indicated also by the work of Gilman and Baet- 
jer on chick embryos^ which showed that exposed hen's eggs 
develop even faster than control eggs for a few days although 
subsequently development is markedly altered and checked. 
One of us, likewise, found that exposure to Roentgen rays re- 
peated frequently throughout the day for several days failed to 
prevent the normal course of development in the eggs of certain 
sea-urchins and teleosts during the peroid of exposure. The 
latent period between exposure to the rays and the development 
of a burn is well known to clinicians. 

1 "Ueber den Einfluss der Roentgenstrahlen auf die Haut in gesunden und 
kranken Zustande," Archiv f. Dermatologie und Syphilis, LIX, pp. 87, 241, 421, 

^ Some effects of the Roentgen rays on the development of embryos, Amer- 
ican Journal of Physiology, X, p-. 222, 1904. 

InJiibitive Action of Roentgen Rays. 195 

While the effect of the Roentgen rays is seen chiefly in the 
inhibition or alteration of reproductive activity in the cells of 
animal tissues it is improbable that it is limited to the results 
of such action. Schaudinn has shown^ that individuals of sev- 
eral species of protozoa may be killed by exposure to the Roentgen 
rays for a few hours. Other forms are not, however, thus sus- 

1 Archiv f. die gesammte Physiologic, LXXVII, p. 29, 1899. 

2 Schwarz in a recent interesting paper (Uber die Wirknng der Radium- 
strahlen, Archiv. f. die gesammte Physiologic C, 532, 1903) concludes that the 
action of radium rays is due to a decomposition similar to that of a dry dis- 
tillation brought about in albunienoid bodies of the cell. He explains their 
effect on rapidly growing tissues as due to their special power to decompose 

n ^ 





I. Introduction. Methods. 

II. Preliminary Notes on tlie normal Development. 

III. The Development of isolated Blastomeres. 

A. Analysis of the first Quartet. 

(i) General development of isolated micromeres of the first quar- 
tet (la, lb, etc.). 

(2) The primary trochoblasts and their products (i^, i^.i^ 12.1.1^ 


(3) Development of the i^ cells (1/16). 

(4) Development of isolated apical cells and secondary trocho- 


(5) Development of the isolated entire first quartet. 

(6) Summary on the first quartet. 

B. Experiments on Cells of the lower Hemisphere. 

(i) The isolated i/8-macromere (lA, iB, etc.). 

(2) The isolated i/i6-macromere (2A, 2B, etc.). 

(3) The isolated second quartet-cells (2a, 2b, etc.). 

(4) Isolated cells obtained by maceration en masse. 

(5) Summary on the lower hemisphere. 

C. Development of isolated 1/2 and 1/4 blastomeres. 

(i) The partial cleavage in Doitalhim. 

(2) The partial cleavage in Patella. 

(3) The half and quarter larvae. 

IV. Summary. 

V. Discussion of Results. 

198 Edmund B. fFilson. 

The first of these studies {Joiirn. Exp. Zoology, I, i, 1904) 
was especially concerned with the question of cytoplasmic ^re- 
localization in the unsegmented molluscan egg, and gave only 
an incidental account of experiments on the cleavage. In that 
paper both cytological and experimental evidence was presented 
to show that the Dentaliiim egg contains from the beginning defi- 
nitely specified regions, consisting of visibly different materials, 
which stand in such a relation to the morphogenic process that 
the removal of particular areas of the unsegmented egg produces 
corresponding definite defects in the resulting larva. It was 
shown, further, that during the cleavage process these materials 
are definitely distributed to the blastomeres of the early embryo, 
and that when these blastomeres are isolated they give rise al- 
ways to defective larvae, showing the same general character as 
those derived from the corresponding regions of the unsegmented 
egg. I therefore concluded that the development of these eggs 
sustains His's theory of germinal prelocalization ("Organbil- 
dende Keimbezirke") as applied to the unsegmented egg, and 
Roux's mosaic theory as applied to the cleavage process, and is in 
harmony with the theory of formative stuffs. 

In that paper, the evidence for the mosaic character of the 
cleavage was given only in part, including only a brief account 
of the general development, in Dentaliiim, of isolated blastomeres 
from the 2-cell and 4-cell embryos, and of isolated micromeres 
of the first quartet. The present paper offers more detailed evi- 
dence in the same direction, derived mainly from experiments on 
Patella ca^nilea. The comparative ease and certainty with 
which blastomeres of any desired stage may be obtained by means 
of Herbst's calcium-free sea-water led me to hope that a fairly 
complete experimental analysis of the potencies of the cleavage 
cells might be carried out; and I do not doubt that in time such 
an analysis can be effected. Various practical diflRculties, how- 
ever, have rendered the analysis here offered incomplete in sev- 
eral directions. Nevertheless the positive results attained form 
the most detailed and, as I think, convincing evidence of mosaic 
development thus far produced, and in my judgment clearly dem- 
onstrate this general principle in the molluscan egg. Despite the 

Experimental Studies en Germinal Localization. 199 

obvious gaps that they show In some directions, I therefore pub- 
lish the results as they stand, with the hope that they may be 
extended hereafter/ 


The eggs ,of Patella cocrulea were obtained in a mature state from March 
until June, those of Dentalium entalis during June and July. Artificial fertiliza- 
tion is easily effected in Dentalium, but is much more difficult in Patella. In 
the latter case I found, after many trials, that the eggs fertilized more readily 
if first placed for half an hour in sea-water rendered slightly alkaline by the 
addition of 4-6 drops of a 5% solution of potassium or sodium hydrate to 
half a litre of sea-water (the slight precipitate first formed quickly dissolves 
upon agitation). The spermatozoa were also placed in the alkalized water 
for the same length of tiime. From 15 to 20 minutes after fertilization (in the 
same water) the eggs were as a rule transferred to a large quantity of pure sea- 
water brought from the open sea. 

In both forms the opaque tgg is at first surrounded by a very distinct 
membrane, whrich, in the case of the ripe eggs, disappears as the eggs lie in 
water before fertilization, in Patella by gradually dissolving away and disinte- 
grating at several points, in Dentalium by suddenly bursting and being thrown 
ofif. Double fertilization occurs rarely in Dentalium, but very frequently in 
Patella, so that ^in the latter case it is essential to pick out the normal eggs one 
by one with a pipette at the 2-cell stage. In both forms the blastomeres can 
be separated with the greatest ease by means of Herbst's calcium-free sea- 
water — indeed, the action is so energetic that better results are obtained if 
it is restrained somewhat by mixing the artificial water with a certain amount 
of normal sea-water. The eggs were placed in the artificial water shortly 
after both polar bodies had formed, and after division the blastomeres were 
carefully separated under the lens with a fine scalpel and immediately isolated 
in normal sea-water. Even so, however, the blastomeres often continue to 
separate in the normal water, and the best results for the earlier stages were 
obtained by not employing the artificial water, but by separating the cells 
with the scalpel in normal water. This is difficult in Patella, but very easy in 
Dentalium. in the earlier stages. For somewhat later stages the artificial water 
must be used; but this can often be successfully accomplished by transferring 
the 2-cell stages to normal water and separating the blastomeres at the proper 
stage. The tendency to separate after transference to normal water steadily 
decreases as the development proceeds; hence good results for still later stages 
are obtained by allowing the eggs to segment in the artificial water up to the 
16-32-64-cell stages, before isolation and transfer to normal water. For greater 
certainty of identification the best plan is to separate and isolate the blasto- 

iLike the preceding work, this was done at the Naples Zoological Station 
between February and the end of July, 1903, on a grant from the Carnegie In 
stitution of Washington. I would again express my great indebtedness to 
the administration of the Station for the unremitting care and efficiency with 
which my work was aided. 

200 Edmund B. JVilson. 

meres after each division, transfering them to normal water at the stage de- 
sired and all my critical results have been thus attained. The mortality is very 
large, since the blastomeres seem to suffer severely in the change from the ar- 
tificial to the normal water, and is greatest in cells from the vegetable hemi- 
sphere; hence my failure thus far to isolate successfully the second somatoblast 
(or primary mesoblast-cell, 4d), in some respects the most interesting of all 
the cells. For the latest stages I did not endeavor to isolate the cells at all, 
but allowed the eggs to develop for 24 hours in the artificial water, from time 
to time separating the cells by jets from a fine pipette. 

Most of the studies on isolated blastomeres were made on Patella, since 
with Dentalimn most of my time was given to experiments on egg-fragments. 
For preparation of the Patella eggs I found no better method than the simple 
one employed by Patten ('85) ot acetic acid and glycerine. The eggs were 
placed in a watch-glass nearly filled with sea-water, two to four drops of glacial 
acetic acid added, followed by successive additions of dilute glycerine gradu- 
ally replaced with strong glycerine. This renders the embryos perfectly trans- 
parent, with sharply marked cell-boundaries, and often gives preparations of 
admirable clearness. A slight stain with acetic carmine often adds consider- 
ably to the effectiveness of the preparation for a time, though they subse- 
quently deteriorate, and for most purposes the stain is superfluous. 



Unfortunately the cell-lineage of neither Patella nor Dental- 
iiim has been worked out. Patten's early paper on the embry- 
ology of Patella ('85), excellent as it is in many respects, leaves 
this part of the development nearly untouched, and the same is 
true of the still earlier paper of Lacaze-Duthiers ('57) and that 
of Kowalewsky ('83) on Dentalium. Everyone familiar with 
work of this type will appreciate the fact that to work out the 
cell-lineage fully would require prolonged study, and both the 
forms here dealt with present peculiar difficulties in the later 
stages. The time at my disposal has only allowed me to deter- 
mine the main outlines of the cell-lineage, including details es- 
sential to the interpretation of the more important experimental 
results. Fortunately, however, Robert ('02) has recently pub- 
lished a detailed study of the cell-lineage of TrocJiuSj which agrees 
so closely with that of Patella that it may be taken as a standard 

Experimental Studies on Germinal Localization. 201 

of comparison. To facilitate the comparison I shall employ 
Robert's nomenclature, which combines certain advantageous 
modifications, suggested by Conklln, Mead and Child, of the sys- 
tem I used In 1892 in describing the cell-lineage of Nereis. The 
primary quadrants are designated as A, B, C and D (D being the 
posterior one), the corresponding micromeres as a, b, c and d; 
the coefficient (i, 2, 3 or 4) designates the number of the quar- 
tet, or In case of the basals (macromeres) the number of divi- 
sions they have undergone; each exponent denotes a subsequent 
division, i designating the cell nearer the animal pole, 2 the sis- 
ter-cell nearer the "lower pole. Thus, starting with the 4-cell 
stage, D divides Into iD and id; iD into 2D below and 2d 
above; id into id^ above and id' below (the primary trocho- 
blast) ; id^ Into id^-^ above (primary rosette-cell at the upper 
pole) and id^, •- below (primary cross-cell) ; 2D Into 3D and 3d; 
2d Into 2d^ and 2d", etc. Since in Patella the quadrants cannot 
be distinguished by simple inspection before the 32-cell stage I 
shall in general, where the quadrant is unknown, omit the letter. 
Thus the primary trochoblast is i", the primary rosette-cell i^-\ 
a primary quartet-cell i, 2, 3 or 4, and so on. 

Both Patella and Dentalium are typical examples of the spiral 
type of cleavage, the former being of the symmetrical type (like 
Crepidiila, Trochus, Hydroides or Polygordius) In which the 
four quadrants are of nearly or quite equal size, the latter of the 
asymmetrical type (like Nassa, Ilyanassa, Unio, Nereis or Jm- 
phitrite) In which the first division Is unequal and the posterior 
quadrant Is larger than the others until after both have been 
formed. Dentalium, further. Is characterized by the formation 
during the first three cleavages of a large polar lobe which after- 
wards fuses with the posterior cell, CD, the egg passing at the 
first cleavage through the characteristic "trefoil stage" that so 
commonly occurs among mollusks (Nassa, Ostrea, etc.) and oc- 
casionally In annelids (Myzostoma, Sahellaria, Chaetopterus) . 
In a preceding paper ('04) I have sketched the early cleavage of 
Dentalium and will here describe primarily that of Patella. 

The egg of Patella first divides Into equal quadrants, without 
the formation of a polar lobe; and the 4-cell stage Is remarkable 


Edmund B. inison. 

Fig. I. 

Experimental Studies on Germinal Localization. 203 

Normal Development of Patella. 

(From Acetic-Glycerine Preparations; x20o). 

I, 4-celI stage, from upper pole; 2, 8-cell stage, from upper pole, prep::ring 
for fourth cleavage; 3, i6-cell stage, from the side (primary trochoblasts 
shaded); 4, 32-cell stage, from the side; 5, 48-cell-stage (transitional to 56- 
cell stage); 6, 48-cell stage (transitional to 52-cell), from upper pole; 7, 58- 
cell-stage, from upper pole, after division of the rosette-cells and establish- 
ment of the primary cross: 8, ctenophore-stage, about 10 hours, from upper 
pole, primary trocboblasts ciliated; 9, 52-cell stage, from lower pole; 10, em- 
bryo of about II hours, from the right side, showing three secondary trocho- 
blasts in the lateral gap; 11, the same embryo from the left side; 12, portion 
of the same, anterior view; at the opposite end are two secondary trochoblasts 
(primary trochoblasts stippled, secondary unshaded). 

204 Edmund B. Wilson. 

from the fact that in the early stages it often shows no cross- 
furrow (thus differing from Trochiis) ^ or if one is present it is 
very short (Fig. i), though in the 32-cell stage a characteristic 
cross-furrow is present at the lower pole (Fig. 13). It is there- 
fore impossible to identify the quadrants in the earlier stages 
without having observed the divisions from the beginning. As 
usual, three quartets of ectomeres are successively formed by al- 
ternating dexiotropic and leiotropic spiral or oblique divisions. 
The micromeres of the first quartet, often only slightly displaced 
towards the left, are considerably smaller than the basal cells, 
but are relatively larger than in Trochiis (Fig. 2). The fourth 
cleavage is closely similar to that of Trochiis^ each of the upper 
cells dividing slightly unequally and each of the basals somewhat 
more unequally to form the second quartet. In the i6-cell stage 
(Fig. 3) the egg consists as usual of four large basal cells (2A, 
2B, 2C, 2D), four smaller upper cells (la^ — id^) and eight 
alternating cells surrounding the equatorial region. Four of 
these, of equal size, form the second quartet (2a — 2d). The al- 
ternating four, which are somewhat smaller (la- — id"), are the 
primary trochoblasts,^ by two successive equal divisions of which 
arise the sixteen cells of the primary prototroch. The i6-cell stage 
is thus closely similar to that of Trochus, except that the basal 
cells are relatively smaller while all the others are relatively 

The fifth cleavages are dexiotropic and symmetrical through- 
out the embryo, and again agree in the main with those of Tro- 
chus. Each of the basals divides unequally to form a cell of the 
third quartet, relatively somewhat larger than in Trochiis^ while 
each cell of the second quartet divides nearly equally (in Trochus 
this division is distinctly unequal). The primary trochoblasts di- 
vide equally, the upper cells unequally, so as to form at the upper 
pole a rosette of smaller cells (Fig. 6) almost identical with those 
in Trochus, but slightly larger. 

The 32-cell stage thus attained (Fig. 4) is at first perfectly 
radially (spirally) symmetrical. From the four large symmet- 

^These cells, and their products, are stippled in all of the figures 

Experimental Studies on Germinal Localizaton. 205 

rically placed basal cells arises the ento-mesoblast, while as usual 
the 28 remaining cells constitute the ectoblast. 

The sixth cleavages (32-64 cells) are in the main oblique and 
leiotropic; but unlike Trochus the posterior micromeres of the 
third quartet depart more or less widely from the type. Proceed- 
ing from the upper pole downwards the divisions are as follows. 
The rosette-cells (i/"^) divide nearly equally in regular spiral 
order exactly as in Trochus^ so as to form a symmetrical group of 
eight small cells at the upper pole (Fig. 7) which form, certainly 
in part and probably as a whole, the basis of the apical organ. 
Nearly at the same time the i^- cells divide nearly equally, so as 
to form the primary "cross," which, as in Trochus, has at this 
period spirally curved arms (Figs. 6, 7). The trochoblast-pairs 
(i"-^ and I'-) divide equally, somewhat earlier than the fore- 
going, so as to produce four symmetrically placed groups of four 
equal cells (Figs. 5-7). This division takes place much earlier 
than in Trochus, and no further division occurs In the products, 
which become ciliated from the eighth to the tenth hour and form 
the primary prototroch. The second quartet cells divide at about 
the same time in a very characteristic fashion that is almost identi- 
cal with that occuring In the nemertine egg and nearly similar to 
that of Trochus. The upper left cell (2.^) divides slightly un- 
equally, the smaller cell lying above and between the two ad- 
joining trochoblast groups (Fig. 5). The lower right cell (2-) 
divides still more unequally, the smaller lower cell (2--) lying 
below against the corresponding macromere, and between the two 
adjoining cells of the third quartet. In Trochus this cell Is smaller 
still. The egg thus attains a 56-cell stage, at which a slight pause 
occurs, and In the meantime a marked change occurs in one of the 
macromeres which, I think. Is undoubtedly the posterior one, 3D. 
This cell rapidly passes into the Interior, Its outer end becoming 
greatly reduced, and being connected with a narrow neck with 
a swollen interior portion, the nucleus however still lying at the 
surface (Figs. 9, 13). The next cells to divide are those of the 
third quartet. The two anterior ones divide leiotropically, like 
the preceding micromeres. In the two posterior ones, however, 
the spindles assume a bilateral position, with the central poles 

2o6 Edmund B. JFilson. 

close against the outer end of 3D (Figs. 9, 13) ; and while I 
have not actually seen the division, it is nearly certain from the 
position of the spindles that the division is unequal. The study 
of a good many preparations of this stage leads me to believe that 
this is a constant relation. 

The last cells to divide in the sixth cleavage are the macro- 
meres, and of these 3D is the first. At the time of its division it 
is only connected with the surface by a very narrow neck, as- 
suming the extraordinary appearance shown in Fig. 14. The 
result of this division is to form a large rounded cell, that lies 
quite in the upper hemisphere (shown in Figs. 15, 16) and a 
more superficial cell. From the conditions observed at a slightly 
later stage I believe the former to be 4D, the latter the primary 
somatoblast 4d or M ; but I am not entirely certain of this identi- 
fication. Slightly later the remaining macromeres divide some- 
what unequally, the cells in the meantime undergoing consider- 
able shiftings and extending further up into the egg, so that it is 
exceedingly difficult to identify them individually. The ectoblast- 
cap has now extended far down towards the lower pole, so that 
the macromeres are connected with the surface by narrow necks. 
The cell I believe to be 4d now divides symmetrically into two 
to form two large symmetrical cells lying between the entomeres 
and the ectoblast (Figs. 15, 16), which correspond with the 
mesoblast pole-cells as figured by Patten (<?. g., in his Figs. 27, 
36). I have not positively traced these cells into the coelome- 
soblast, but believe there can hardly be a doubt as to their na- 
ture.^ At this period the large inner cell (identified as 4D) is 
still undivided (Fig. 15) the primary trochoblasts have devel- 
oped cilia, and the apical tuft is present (10-12 hours). 

Beyond this point I shall not for the present attempt to trace 
the general cleavage, but will pass on to some points in the later 
development. Patten has given figures of the trochophore of 

^Sections of the trochophores of 24 hours clearly show two large meso- 
blastic pole-cells (one of which appears in Fig. 17) near the posterior end, 
from which two mesoblast-bands extend forward as figured by Patten, c. g., 
in his Fig. 50. 

Experimental Studies on Germinal Loealization. 207 

Patella with which in the main my observations agree, though 
the arrangement of the cells of the prototroch is somewhat 
schematized. The embryo becomes ciliated at about eight 
to ten hours (depending on the temperature) the first cells to 
acquire cilia being the sixteen primary trochoblasts. For a brief 
period the prototroch consists of only those sixteen cells, still ar- 
ranged in four separate groups (Fig. 8). The cilia are from the 
first arranged, not in vague patches or tufts, but in very definite 
oblique transverse rows, which bear a marked resemblance to 
the swimming plates of a ctenophore — indeed, it hardly seems 
forced to compare the embryo directly at this period to a larval 
ctenophore. At the same time, or a little later, the group of 
small cells at the apical pole, derived mainly, if not wholly, from 
the apical rosette, develops a tuft of flexible but non-vibratile sen- 
sory flagella, and constitutes the apical organ. 

The ctenophore-stage is of short duration. In two or three 
hours several cells lying in the gaps between the four groups of 
primary trochoblasts also become ciliated and ultimately enter the 
prototroch as secondary trochoblasts. These cells are not more 
than half the size of the primary trochoblasts (a point of im- 
portance in connection with the experimental results) , and at 
first bear much smaller cilia. There are at least three and prob- 
ably four of these trochoblasts in each quadrant (with the pos- 
sible exception of the posterior group, in which I have only 
certainly seen two of these cells), (Figs. 10-12), giving a total 
of 28 to 32 cells in the prototroch, to which possibly still others 
may be added. While I have not traced step by step the exact 
origin of these cells, their position in the embryo leaves little 
doubt that in each quadrant two of them are derived from the 
first quartet (/. e.^ from derivatives of the i'- cells), and this 
is demonstrated to be the case by the experimental evidence. The 
position of the third cell (Fig. 10) shows almost beyond a doubt 
that it is derived from the second quartet, /'. e., from 2^ and 
probably from 2.^\ The experimental evidence again proves 
that at least one, and in some cases two, trochoblasts are derived 
from the second quartet. As may be seen in Fig. 10, a second 
cell lies next to the one described, the position of which indicates 


Edryiund B. Wilson. 

Fig. II. 

Experimental Studies on Germinal Localization. 209 

Normal Development and Larva from Egg-Fragment, Patella; x 200. 

13, 48-cell stage, lower pole: 14, larva of 9 hours, sagittal optical section, 
division of 3D; 15, larva of 12 hours, optical sagittal section, showing left 
primary mesoblast (?) in division; 16, the same larva, in frontal optical sec- 
tion; 17, trochophore of 30 hours, from the left side (from a total preparation, 
shell-gland (s. g.) and primary mesoblast (m) from a corresponding actual 
section), prototrochal cells in surface-view, body-wall in section; 18. larva of 
20 hours, from upper pole, showing the cells as accurately as possible (some 
of those just anterior to prototroch could not be clearly seen and have been 
omitted); 19, optical section at the level of the prototroch of normal larva 
of 24 hours; 20, larva of 24 hours from fertilized egg-fragment that segmented 
like a whole egg, prototrochal cells in surface-view; 21, optical section of the 
same larva at the level of the prototroch; in all these figures the prototrochal 
cells are shown as acurately as possible. 

2IO Edmund B. ffllson. 

that it may also enter the prototroch.^ The 28 (32?) trocho- 
blasts are at first arranged in two roughly alternating rows en- 
circling the embryo slightly above the equator; and the ciliary 
plates of contiguous cells are still not united to form a continu- 
ous ciliary girdle (Figs. 10-12). Later, extensive shif tings of 
the cells occur in such wise that a principal circle of trochoblasts 
is formed in a single circle completely surrounding the embryo, 
bearing a perfectly continuous series of powerful cilia (Figs. 
17-19). The cells in this row vary in number from 19 to 21 — 
a fact of which no doubt is left by the study especially of acetic- 
glycerine preparations, in which the cells may be seen with sche- 
matic clearness. Posterior to this row lies a second row of smaller 
elongated trochoblasts, which in the dorsal region become as 
large as those of the principal row (Fig. 17). At this point, 
therefore, where in so many trochophores a gap exists in the 
prototroch, the ciliated belt is not only closed, but broader than 
at any other point. At this point the prototroch is often three 
cells wide; elsewhere I have not been able to distinguish three 
rows of cilia as figured by Patten, though three such rows are 
certainly present in Dentalhim. 

The trochophore of 24-30 hours (Fig. 17) is in the main sim- 
ilar to that of Dentalium, as described in my former paper, but 
the post-trochal region is relatively larger, the pre-trochal region 
less pointed, the apical tuft shorter and broader, and the apical 
plate less clearly marked off from the surrounding ectoblast as 
may be very clearly seen in sagittal section. In this respect my 

^This derivation of the prototroch in Patella agrees closely with that of 
Isclmochitou (Heath, 99), where two cells in each quadrant are likewise con- 
tributed from i^-, and two from the second quartet except in the D-quad- 
rant, wbere a non-ciliated dorsal gap exists from the first. I have determined 
beyond doubt, I think, that at least two secondary trochoblasts are formed 
in the mid-dorsal line, as shown in Figs. 10-12, where there are three such 
trochoblasts in three of the quadrants and two in the fourth. There is further 
no doubt whatever that the completed prototroch is closed in the mid-dorsal 
line (Figs. 17-19). Robert describes the prototroch of Troclius as agreeing ex- 
actly with that of Am[>hitrite and Arenicola, no cells being derived from the 
first quartet except the primary trochoblasts. It appears to me, however, that 
his observations do not fully establish this. {Cf. the useful comparative table 
given by Robert at p. 420). 

Experimental Studies on Germinal Localization. 


larvae seem to differ somewhat from those figured by Patten, 
which show an extremely distinct apical plate. The later stages 
are in the main similar to those described by Patten, and need 
not here be considered. 



In general, as Crampton ('96) found for the 2- and 4-cell 
stages of lUyanassa, the isolated blastomere, at whatever stage 
it be separated' from its fellows, continues to segment essentially 
in the same way as if forming a part of a whole embryo; but a 
point on which I would lay stress is that there is a tendency for 
all unequal divisions to be less unequal than in the normal devel- 
opment, though this is by no means always the case, and the 
isolated blastomere often divides exactly as in a whole embryo. 
The partial character of the cleavage is also frequently masked 
by shifting of the cells, and the partial embryos often close, some- 
times at a very early period. Such shifting or closure appears, 
however, to have no effect on the differentiation of the cells, as 
is shown with especial clearness by the history of the trochoblasts. 
Differentiation takes, in the main, the same course as if the cell 
had remained united to its fellows, and gives rise to structures 
that agree in a general way, and sometimes exactly with the parts 
to which the cells would have given rise in a complete embryo. 
For the sake of clearness I shall not follow the most logical or- 
der, but will present first the cases that most completely sustain 
the above statement — namely, the blastomeres of the first quartet. 
It may be premised that all of the isolated blastomeres assume a 
nearly or quite spherical form before division occurs, showing 
no trace of flattening on one side; and they are indistinguishable 
from one another except in size, and in the slightly greater trans- 
parency of the micromeres. It is also necessary to bear in mind 
that both in Dentalium and in Patella the eggs from different fe- 
males vary very considerably in size, so that exactly correspond- 
ing blastomeres from different eggs likewise present consider- 

212 Edmund B. Wilson. 

able size variations. This accounts for certain discrepancies in 
the figures, which represent blastomeres from many different eggs, 
and possibly even from different species, though most of them 
are from P. cceridea. The typical size-relation in this species, 
from the eggs of a single female, are shown in Fig. loo, the 
successive concentric outlines representing the entire egg, the J-2- 
blastomere, M-blastomere, >^-macromere and >^-micromere. Dis- 
tinct deviations from these mean volumes will be observed in the 


I. General development of isolated micromeres of the first 
quartet (^s -embryos) . 

As described in my preceding paper, the development of the 
posterior micromere of this quartet (id) in Dentalium differs 
from that of the others in being the only one to form an apical 
organ. In Patella this is not the case, each micromere giving rise 
to a closed ectoblastic structure, bearing at the posterior end a 
group of active trochoblasts and at the anterior end an apical 
organ (Figs. 28-29).^ 

The first cleavage is slightly unequal (sometimes nearly or 
quite equal), (Figs. 22-24). I ^t first supposed the smaller cell 
to be the primary trochoblast (i") since in the whole embryo 
this cell gives the appearance of being slightly the smaller (in 
Trochus this division is described as "nearly equal"). When, 
however, the entire j4-blastomere segments in the calcium-free 
water it may clearly be seen, at least in some cases, that the larger 
cell is the lower one (i'), and I believe therefore the primary 
trochoblast is slightly larger than its fellow in the normal devel- 
opment. This is typically followed by an equal division of the 
trochoblast, and an unequal division of the other cell, giving a 
group that closely represents one quadrant of the first quartet in 

^This has not been proved in Putclla by isolation of all four of the micro- 
meres from one t^^g (as was done in Dentalium') ; but among the numerous 
larvae obtained of this type all that were closely examined possessed the api- 
cal organ. 

Experimental Studies on Germinal Localization. 213 

3/ ^32 ^33 

Fig. III. 

Development of Isolated ^s Micromeres. 

(Figs. 22-27 >^ 250; -Figs. 28-30 X 290) 

22, isolated micromere; 23. 24, first division; 25, 26, two different indi- 
viduals, 32/8 stage, each with two trochoblasts, one rosette-cell, and one pri- 
mary cross-cell; 27, an entire quadrant segmenting after removal from cal- 
cium-free water, products of first and second quartets somewhat separated 
from the basal (the first quartet group seen from the outside, the others from 
the inside). 28, larva of 24 hours, from J/g-micromere, from the side, showing 
trochoblasts below, apical cells above; 29, similar larva (with less regular pre- 
trochal region), from below, showing both primary and secondary trocho- 
blasts; 30, loose group, from J^-micromere, after 24 hours in water nearly 
free from calcium, primary and secondary trochoblasts, apical cells, pre- 
trochal ectoblast-cells; 31, group, with two apical cells, from ^-micromere. 
after 24 hours in calcium-free water; 32-33, isolated apical cells from similar 

2 14 Edmund B. Ifllson. 

a normal 32-cell stage — z. e., consists of two trochoblasts, one 
rosette cell (i^'^), and its larger sister cell (i^"'), from which 
one arises one arm of the cross (Figs. 25-27). It should be 
noted that the rosette-cell almost always appears somewhat too 
large, which is owing In part to the fact that it is less crowded 
than in a whole embryo, but undoubtedly in part also as to a les- 
sened inequality in the division of i\ Such embryos give rise to 
actively swimming partial larvae, similar in a general way to the 
corresponding ones in DentaJ'uim. These embryos do not gas- 
trulate, but close to form pyriform ectoblastic larvae, which bear 
at the larger end a group of large ciliated trochoblasts, and at 
the narrower end an apical organ consisting of a group of cells 
bearing stiffish motionless sensory hairs (Figs. 28-29). It may 
be clearly seen that the larger end of the embryo is formed of 
four primary trochoblasts, each bearing a row of powerful cilia, 
while just above these at one side are two somewhat smaller sec- 
ondary trochoblasts. The apical organ at this period appears, 
in most cases at least, to include only two cells from which the 
sensory hairs radiate like a fan. This differs from the normal 
apical organ, in which the sensory hairs form a thick tuft directed 
straight forwards. The radiating arrangement of these hairs 
in the partial embryos appears to be due to the fact that the apical 
cells do not extend so deeply below the surface, and retain a 
rounded form, so that the sensory hairs spread apart like a fan, 
while in the normal embryo they are crowded together and as- 
sume a pyramidal shape, the free surface being considerably re- 
duced. In the partial larvae, too, the sensory hairs appear rela- 
tively shorter and more rigid than in the normal organ. 

The composition of these larvae is shown with great clearness 
by allowing the isolated ^s-micromere to develop in calcium-free 
water, the action of which is more or less restrained by the ad- 
dition of a certain amount of normal sea-water. All degrees of 
dissociation may thus be obtained, and among the resulting cell- 
groups may be found forms like Fig. 30, in which the cells lie in 
a loose group, yet approximately retain their normal position. It 
is evident that each of these larvae represents one quadrant of the 
products of the first quartet, including four primary trochoblasts, 

Experimental Studies on Germinal Localization. 215 

two secondary ones, one-fourth of the apical organ and a group 
of small ectoblast cells derived from i^-^, the whole structure 
closing to form a morula or blastula-like structure, but otherwise 
differentiating typically without gastrulating. In the aquarium 
these larvas gradually disintegrate in the course of the second or 
third day, the trochoblasts being always the longest-lived of the 
cells, and often continuing to swim actively when the remainder 
of the larvae has gone to pieces. I have not followed the details 
of the development of the corresponding Isolated cells of Denta- 
lium; but it is clear that their general development is closely simi- 
lar. The one important difference, pointed out above, is that in 
Dentaliiim only the micromere from the D-quadrant develops an 
apical organ. As my experiments on Dentaliiim showed, the de- 
velopment of the apical organ in this form is determined by ma- 
terial that originally lies In the polar lobe, and no other conclu- 
sion seems possible than that this material is in Dentaliiim finally 
isolated In the posterior micromere (id), while in Patella the 
corresponding stuff is equally distributed among the four mlcro- 
meres. This is doubtless due to a different relation. In the two 
cases, of the original segregation pattern to the first two cleavage 
planes, and is perhaps connected with the absence of a polar lobe 
In Patella. 

2. The primary trochoblasts (1-16, 1-32, i-64-embryos) . 

Exceedingly clear and interesting results are given by the Isola- 
tion of the primary trochoblasts (i") or their products. If a 
single trochoblast be isolated at the i6-cell stage It divides equally 
twice in succession, but no further division takes place (Figs. 
34-39) . From the eighth to the tenth hour each of the four cells 
becomes ciliated and the group begins to swim. After twenty- 
four hours the group Is swimming with great activity, and each 
cell is found to bear a series of powerful cilia arranged in a 
transverse row, exactly as In a normal embryo (Figs. 40-42). 
The cells vary In arrangement, sometimes lying in a single plane, 
sometimes having shifted so as to interlock in a rounded mass. 
Exactly as in a normal embryo the action of the cilia is more or 


Edmund B. Wilson. 


Fig. IV. 

Isolated Primary and Secondary Trochoblasts. 

(Figs. 34-39 X 250; Figs. 40-48 x 290). 

34, primary trochoblast (1/16, i-), obtained by successive isolation; 35, 36, 
result of first division; 37-39, various forms after second division; 40, product 
of Figs. 34, 36, 37, after 24 hours; 41, 42, similar individuals of the same age 
and history; 43, pair of primary trochoblasts, 24 hours, the products of i^-i or 
i2.2 (1/32); 44, 45, single primary trochoblasts. 24 hours, products of i^-i-i, 
i2-^-2, etc. (1/64); 46, pair of secondary trochoblasts, 24 hours, the products of 
1^-2; 47, 48, single secondary trochoblasts, 24 hours. 

less intermittent, sometimes ceasing wholly and again suddenly 
being resumed. Sudden mechanical shock frequently causes a 
sudden suspension of activity, followed immediately by a more 
vigorous activity than before. Careful study of these embryos, 
especially when they are dying, shows that the cilia in each cell 
beat in the same direction; but owing to the fact that the rows 

Experimental Studies on Germinal Localization. 217 

of cilia rarely coincide in direction the group does not, as a rule, 
rotate in a constant direction, but irregularly. 

If now the two products (i^-^ and i"-) of the first division of 
the primary trochoblasts be separated, each divides once, and only 
once, thus giving a pair of cells that become ciliated and swim 
together like the above-described group of four (Fig. 43). If, 
finally, these two cells be separated at the time of ciliation — i. e., 
at a period corresponding with the 64-cell stage, no further divi- 
sion occurs, but In due time each trochoblast develops its row 
of cilia (Figs. 44-45) and swims singly with the fullest vigor 
and activity. Such single trochoblasts often rotate steadily in a 
nearly constant direction, proving that the action of the cilia 
is normally coordinated. They may live for two days or more 
when the action gradually ceases and disintegration occurs. 

The history of these cells gives indubitable evidence that they 
possess within themselves all the factors that determine the form 
and rhythm of cleavage, and the characteristic and complex dif- 
ferentiation that they undergo, wholly independently of their re- 
lation to the remainder of the embryo. Roux's "self-differentia- 
tion" here appears in the clearest and most unmistakable form. 

Similar results were obtained in Dentalium, but I did not in this 
case attempt to isolate the trochoblasts individually, but merely 
allowed entire eggs, or isolated ^ or ^-blastomeres to continue 
their development In the calcium-free water. As in Patella, the 
result at the end of 24 hours is a chaos of more or less com- 
pletely separated cells of different forms and sizes, among which 
are trochoblasts, actively swimming, singly or In groups. These 
trochoblasts fall roughly into three groups, large (Figs. 49-51), 
medium (Figs. 52-53) and small (Figs. 54-55). The large 
trochoblasts, which are considerably larger than in Patella, are 
probably primary ones, the medium and small forms secondary 
ones; and the difference in size among the latter suggest that as 
in Patella they may arise from different quartets. It is worthy 
of note that the cilia in all the trochoblasts are considerably longer 
than in Patella, and are also relatively less numerous and 
crowded. The small trochoblasts sometimes have as few as six 
cilia (Fig. 55). Sections of the entire larvae show that the cilia 


Edmund B. Wilson. 

Fig. V. 

Isolated Troclioblasts of Dcntalium and Mcsoblast-likc Cells of Patella 24 Hours: 

X 290. 

(All these obtained by leaving entire embryos in the calcium-free water for 
24 hours). 

49-51, large (primary) trochoblasts; 52, 53, medium (secondary?) trocho- 
blasts; 54, 55, small (secondary?) trochoblasts; 56, group of mesenchyme-Hke 
cells; 57, group of muscle-like cells. 

Experimental Studies on Germinal Localization. 219 

are grouped in small tufts (an arrangement I failed to note in 
the isolated cells) at the base of each of which is a very distinct 
deeply staining basal body; and I believe this would be an ex- 
cellent object for the cytological study of the possible relation of 
these bodies to the centrosome. 

3. Development of the sister-cell (i^) of the primary trocho- 
blast ( i/i6-embryos) . 

The development of this cell, which, except for its slightly 
smaller size, is indistinguishable in appearance from the primary 
trochoblast, differs totally from the foregoing in the form and 
rhythm of cleavage, and in the course of its differentiation. After 
isolation this cell typically divides unequally to form a single 
rosette cell (i^-^) (though here, too, the inequality is often less 
marked than in the normal embryo, and sometimes disappears), 
and its larger sister (i^"), (Figs. 58-60) ; and this is typically fol- 
lowed by a nearly equal division of both these cells to form a 
group of four, two of which obviously represent daughter-rosette 
cells (Figs. 61, 62). The divisions do not cease here, however, 
but continue, and at the end of 24 hours a larva is produced that 
consists of many cells and is somewhat similar to those arising 
from an entire micromere of the first quartet; this larva is, how- 
ever, only about half as large, and lacks the four large trocho- 
blasts at the posterior end (Figs. 63-64). At this end of the 
embryo are two trochoblasts, much smaller than those of the pri- 
mary group, which obviously represent the secondary trocho- 
blasts of the first quartet. At the narrower anterior end is an 
apical organ precisely like that of the micromere j/^-embryo, 
while the middle region consists of small ectoblast cells usually 
larger on one side than on the other. These embryos swim ac- 
tively, but less vigorously than the J^-forms, and, like the lat- 
ter, perish in the course of the second or third day. It is obvious 
that the development of the i^ cell is, except for its closure, es- 
sentially the same when isolated as when it forms part of a whole 
embryo; and its remarkable contrast with its sister-cell, the pri- 
mary trochoblast, shows most convincingly that despite their 


Edmund B. JVilson. 

Fig. VI. 

Isolated i^ Cells and Isolated First Quartet, Patella. 
(Fig-s. 58-62, 65, 66 X 250; Figs. 63, 64 X 290). 

58, isolated iM 59, 60, two examples of first division (rosette-cell slightly 
too large in both); 61, 62, two examples of second division; C>3, product, 12 
hours, apical cells and secondary trochoblasts (somewhat smaller cells on re- 
verse side); 64, similar larva of 22 hours; 65, first division of isolated first quar- 
tet; 66, second division; 67, product, 24 hours. (Some of the cells have been 

Experimental Studies on Germinal Localization. 221 

closely similar external appearance, each is from its first forma- 
tion definitely specified, irrespective of its connection with its fel- 

4. Development of isolated apical cells, and secondary trocho- 
b lasts. 

In spite of many attempts, I did not succeed in rearing singly 
one of the rosette cells; but isolated products of these cells, as 
well as isolated secondary trochoblasts, were obtained in another 
way. This was by allowing isolated micromeres of the first quar- 
tet to continue their development in the calcium-free water, some 
of the individuals being left undisturbed, others shaken to pieces 
from time to time by a stream from a fine pipette. In those left 
undisturbed for 24 hours all degrees of disintegration were ob- 
served, loose masses being found from which the trochoblasts 
often had broken away and were swimming about singly. In 
many such loose masses the characteristic apical cells could often 
be observed, loosely attached to their fellows at the end opposite 
the trochoblasts (as in Fig. 30). Those that had been shaken 
to pieces showed a collection of more or less completely separated 
rounded cells, among which spherical trochoblasts of two sizes 
(the larger evidently being the primary, the smaller (Figs. 46- 
48), the secondary ones) were actively swimming singly or in 
groups. Among these cells occur forms that are evidently sin- 
gle apical cells, since they agree exactly in size and structure with 
those attached to the loose masses referred to above. These cells 
(Figs. 31-33) are ovoidal in form, and bear on one side the 
characteristic non-vibratile radiating sensory processes or hairs. 
There is no possibility of mistaking these cells for either kind of 
trochoblast, since they are considerably smaller, and the appear- 
ance and arrangement of the sensory hairs (apart from their im- 
mobility) is entirely different from that of the cilia. There can, 
therefore, be no doubt that both the secondary trochoblasts and 
typical sensory cells of the apical plate may undergo their charac- 
teristic differentiation when entirely isolated fro?fi their fellows. 
Mingled with the foregoing cells are rounded, non-ciliated cells 

22 2 Edmund B. JVilson. 

of various sizes that are obviously isolated cells of the general 
pras-trochal ectoblast. 

The relatively large size of the apical cells, whether completely 
isolated or forming part of the ]'i or ^o larvae, is a matter I 
have not yet fully cleared up. The apical pole of the normal 
larva of 24 hours (Fig. 18), seen in surface view, gives some- 
what varying appearances, but clearly shows a central group of 
four to six larger cells. This does not exactly agree with what 
appears in the partial larvae, whiqh, as a rule, show two large 
apical cells, apparently of nearly equal size; it is, however, dif- 
ficult to determine the exact size of the cells in the normal larva, 
owing to their crowding together. It seems probable that this ap- 
parent discrepancy may be due to the fact that, as stated above, 
the primary rosette cell is so frequently too large in the }i and 
tV embryos, and that this results in a slightly abnormal later 
development of its products. It is clear, however, that this does 
not affect in any essential way the differentiation of the apical 

5. Development of the isolated entire first quartet. 

In the sea-urchin Driesch (1902) has shown that both the 
upper and the lower quartets of the 8-cell stage may produce 
complete dwarf larvae, though the two quartets show certain char- 
acteristic differences in development, proving that they are not 
identical. In Patella it is difficult to perform this operation, and 
still more difficult to rear the larvae, since the cells always separ- 
ate more or less after replacement in normal water. I neverthe- 
less succeeded in obtaining a few cases. The cleavage of the iso- 
lated first quartet is essentially the same as in a whole embryo. 
The first division, leiotropic in all the cells, produces four upper 
cells (i^) and the four trochoblasts (i") alternating with them, 
the eight cells forming a nearly flat plate (Fig. 65). The sec- 
ond division, dexiotropic in all the cells, produces a plate of 16 
cells (Fig. 66) in the centre of which is the rosette, around which 
lie the four i^-^ cells and, at the margin, four groups each of two 
trochoblasts (i^S i'"). As shown in the figures, the form and 

Experimental Studies on Germinal Localization. 223 

grouping of these cells is essentially the same as in the upper half 
of a normal 32-cell stage; though, owing to the flattening out 
of the group, the normal position of the cells is somewhat modi- 
fied and (as in the case figured) the cells are often rather loosely 

In later stages such groups invariably broke up more or less, 
and no larvae were obtained which had not lost some of the cells. 
Nevertheless these larvae close up more or less completelv, form- 
ing irregularly pyriform structures with an apical organ at the 
smaller end and a group of trochoblasts at the larger one. The 
largest of these larvae obtained probably represents at least three- 
fourths of the first quartet. This individual (24 hours) is shown 
in Fig. 67, drawn from a preparation; in life it swam very ac- 
tiv^ely about. This larva is clearly a purely ectoblastic structure, 
and shows no trace of archenteron. It is of an irregular flattened 
pyramidal form, with an irregular group of apical cells at the 
narrow end, forming an unmistakable apical organ. The larger 
end is occupied by a group of trochoblasts, which form a some- 
what irregular series around the margin, but also extend some- 
what over the base. The remainder of the embryo is formed of 
small ectoblast cells that have on the lower side extended more 
or less into the basal region. The exact number of trochoblasts 
cannot be determined, but there are at least 14, and probably a 
larger number. It is clear that this larva represents a partly 
closed and distorted prae-trochal region, with that part of the 
prototroch derived from the first quartet, minus a certain number 
of cells that have separated. The other larvce were similar in 
type, but evidently represent a smaller portion of the same 
region. This case, taken In connection with the other facts de- 
termined, renders it practically certain that the first quartet as a 
whole is here incapable of producing a complete dwarf, but gives 
rise to essentially the same ectoblastic structures as in a whole 
embryo. This result is entirely in agreement with that after- 
wards obtained on Cerehratiiliis by Zeleny ('04), who, at my sug- 
gestion, undertook a comparison in this form of the upper and 
lower quartets of the 8-cell stage — a question particularly inter- 
esting in this case since the upper quartet is larger than the lower. 

2 24 Edmund B. Wilson. 

The constant result of this experiment was, that while both quar- 
tets produce closed hlastulas, only the lower one gastrulates, while 
only the upper one produces an apical organ. 

6. Summary on the first quartet. 

The foregoing observations are sufficient, I believe, to establish 
definitely the mosaic character of cleavage and differentiation in 
the first quartet. This is strictly proved for the primary trocbo- 
blasts (i^), for their first (i"-^ and i'-) and second (I'-^'S I^•^■^ 
etc.) products, and for the i^ cells; it is less strictly but hardly less 
convincingly established for the apical cells and the secondary 
trochoblasts. Excepting the secondary trochoblasts, the other 
products of i^ do not show sufficiently definite characters to allow 
of a similar definite proof, but it can hardly be doubted that the 
same conclusion applies to them, and also to the first quartet taken 
as a whole. 


Isolated blastomeres of the lower hemisphere are in general 
much less tenacious of life than those of the upper quartet (a fact 
parallel to that observed by Driesch in sea-urchins), and my ob- 
servations are here much less detailed. In a more general way, 
however, the results are entirely in agreement with the conclu- 
sions reached in case of the upper quartet. 

I. Development of the isolated Ys-macromere. 

These cells divide, at least as far as the 64/8-cell stage, as if 
forming part of a complete embryo. At the first division a cell 
of the second quartet is formed (Fig. 69-70), which at the en- 
suing cleavage divides nearly equally (into 2^ and 2-), while a 
cell of the third quartet is produced from the basal in the proper 
position (Fig. 71). At the next cleavage 2^ divides nearly 
equally, 2" very unequally, forming below the characteristic small 
cell 2--, that lies against 3. The latter cell then divides equally 
or unequally (the quadrant probably being in the former case one 

Experimental Studies on Germinal Localization. 225 

of the anterior, in the latter case one of the posterior ones). A 
group is thus produced (Figs. 72-73) which, excepting the usually 
lessened inequality of 2^-^ and 2^', is practically identical with one 
quadrant of the lower hemisphere in the 64-cell stage (Cf. Figs. 
4-5). This is followed by a division, sometimes distinctly un- 
equal, sometimes nearly equal, of the basal to form a cell of the 
fourth quartet (Fig. 73). 

Fig. VII. 

Isolated 1/8 Basal, Patella; x 250. 

68, isolated basal (a rather small example; cf. Fig. 100); 69, 70, examples 
of first division; 71, 32/4-cell stage, typical; 72, 56/8-cell stage, typical 2- 
group, equal divison of 3; 73, 64/8-cell stage, after formation of 4 (in this ex- 
ample the 2-group has rotated into an abnormal position); 74, product, 24 
hours, showing two secondary trochoblasts (products of 2i-i) and two feebly 
ciliated cells (pre-anal cells?). 

226 Edmund B. fnison. 

It is exceedingly difficult to rear these embryos, many of them 
dying, while most of those that live break up into smaller masses. 
A few larvae were nevertheless obtained, the best of which is 
shown in Fig. 74. Like the others, this larva has evidently gas- 
trulated (though the entoblast-mass is relatively small, perhaps 
owing to the loss of some of the cells). Its most interesting fea- 
tures are the presen-ce, at one end, of two cells bearing powerful 
and active cilia by the activity of which the larva rotates irregu- 
larly, while near the opposite end are two cells bearing much 
smaller and feebler ones. It is evident that the two anterior cells 
are secondary trochoblasts, undoubtedly those derived from the 
second quartet; and the fact that there are two of these may be 
taken as evidence that two trochoblasts are contributed to the 
prototroch, at least in some of the quadrants, by the 2^-^ cells, as 
is indicated by a study of the normal embryos. The two weakly 
ciliated cells were to me at first a puzzle, since I failed to observe 
anything corresponding to them in the normal embryos. But it 
may be recalled that Patten describes and figures a ventral reg- 
ion, covered with fine short cilia, just anterior to the pre-anal 
sense-organ ('85, Figs. 47, 48, 57, etc.), and while I became 
aware of this when it was too late to re-examine the normal larvae, 
it seems very probable that it is these cells that appear in the 
posterior ciliated tract of the >^-macromere larva. This is sus- 
tained by the development of the yV basal cells about to be de- 

2. Development of the i/i6-macromere. 

The tV basal cell, obtained by successive isolations, divides 
unequally to form the third quartet cell (Fig. 75), which after- 
wards divides into two, while still later the fourth quartet-cell is 
produced from the basal (Fig. 76). Only two or three such 
cases were obtained, one of which developed into the larva shown 
in Fig. 77. While this larva could not be very clearly analyzed, 
it evidently consisted of an internal mass of cells (entomesoblast 
or entoblast) surrounded by a superficial layer of ectoblast-cells. 
At one end the ectoblast-cells were larger and at least one of 

Experimental Studies on Germinal Localization. 227 

these bore a tuft of short, weak ciHa, by means of which the 
larva very slowly rotated. In view of the structure of the Ys- 
macromere larva, it is probable that this ciliated cell (or cells) 
represents at least a part of the posterior group of cilia which I 
have assumed to contribute to the ventral ciliated tract in the nor- 
mal larva. 

Fig. VIII. 

(Figs. 75-81 X 250; Fig. 82 X 300). 

Isolated 1/16 Basal, and Isolated Cell of Second Quartet, Patella. 

75, first division of 1/16 basal; "jG, second division (64/16), 4 formed; 77, 
product, 24 hours; 78-81, cleavage of isolated 2; 82, product, 24 hours, with two 
trochoblasts and pre-anal (?) cells. 

3. Isolated blastomeres of the second quartet ( i/i6-embryo) . 

These blastomeres, obtained by successive isolations, divide like 
the foregoing as if still forming part of a complete embryo. The 
first division is nearly or quite equal. In the second division one 
of the cells divides nearly equally, the other very unequally. Thus 
arises a 64/16-stage (Figs. 78-81) that is closely similar to the 

228 Edmund B. Wilson. 

corresponding group in a whole embryo (Cf. Fig. 5), though 
these divisions are often (as in all the foregoing cases) less un- 
equal than in a whole embryo, and their arrangement is fre- 
quently modified by shifting of the cells. At the end of 24 hours 
these groups produce closed ovoidal or irregular ectoblastic vesi- 
cles that swim rather slowly by means of a tuft of cilia at one 
end. In some cases these cilia seem to be borne by a single cell; 
in others I am sure there are two of these cells (Fig. 82). These 
cells are evidently secondary trochoblasts; and their presence is 
entirely in agreement with the facts observed in the J^-macro- 
mere-larva described under ( i ) , and with the fact that the nor- 
mal larva clearly shows the derivation of at least one, and prob- 
ably two, secondary trochoblasts from the second quartet. 

There are two additional noteworthy points in these larvse. 
One is the fact that, in addition to the one or two secondary 
trochoblasts, some of them, at least, show one or two other small 
patches of short and feeble cilia like those seen in the }i or tV 
macromere larva. If my interpretation of these cells is correct, 
this may be taken as evidence that the ventral ciliated tract arises 
from derivatives of both the second and third quartets. 

A noteworthy point in these embryos is the presence, in some 
of them, of loose groups of rounded cells lying within the cavity 
(Fig. 82). These cells, considerably smaller than the entoblast- 
cells, are not improbably mesenchyme cells of the "larval mesen- 
chyme" (paedomesoblast or ectomesobast) ; and it may here be 
recalled that in Crepidiila, according to Conklin ('97), the ec- 
tomesoblast is derived from the second quartet. Soon after the 
stage described the embryos died and disintegrated without fur- 
ther noticeable change, 

4. Observations on isolated cells obtained from larvae that 
have developed' continuously in calcium-free water. 

Beyond the facts recorded above, I have not traced the devel- 
opment of isolated blastomeres from the lower hemisphere. Sev- 
eral times I succeeded by successive isolation in separating single 
cells of the third and fourth quartets, and of the corresponding 

Experimental Studies on Germinal Localization. 229 

basals; but in every case the embryos became abnormal or died 
without division. I therefore resorted to the method of allow- 
ing the eggs to continue their development for 24 hours in the 
calcium-free water, separating the cells from time to time by 
directing a rather strong jet of water upon them by means 
of a fine pipette. In this way the cells may be almost com- 
pletely separated so as to produce what is in effect a pro- 
gressive maceration of the larva without killing the cells. The 
result is most striking. At the end of 24 hours the whole em- 
bryo is disintegrated into its constituent cells, some of them lying 
in small groups, but in favorable cases many are completely iso- 
lated. The greater number of these cells are motionless and 
perfectly spherical, of many different sizes, and still appear to be 
living and in a healthy condition. Among these are swimming 
with great vigor numbers of trochoblasts, singly, in pairs, or 
sometimes in groups of four or three (Figs. 49-55). Measure- 
ments of these trochoblasts show that in Patella they are of two 
sizes, in Dentaliinn, as pointed out above, of three, the larger one 
agreeing perfectly with the primary trochoblasts obtained by in- 
dividual isolation, the smaller with the secondary trochoblasts. 
Here and there can sometimes be seen a single apical cell, with 
its chai'acteristic radiating sensory hairs. 

Among the motionless rounded cells it is impossible to dis- 
tinguish the different categories by their structure, since all have 
the same form and all are filled with yolk spheres. In view of 
the foregoing results, however, it can hardly be doubted that the 
largest ones are isolated entoblast-cells. But the most interest- 
ing cells are those which are not rounded but of a different form. 
Two kinds of such cells can be distinguished, both in Patella and 
in Dentaliinn, namely, spindle-shaped cells (Fig. 57), and branch- 
ing mesenchyme-like cells (Fig. 56). The cells of both forms, 
are relatively small, less heavily laden with yolk, and more trans- 
parent than the others. /;/ all these respects these cells are closely 
similar to the mesoblast-cells, as seen in total preparations or sec- 
tions of the normal trochophore of the same age. 

These facts must be interpreted with considerable reserve; for 
it is well known that isolated cleavage-cells often become irregu- 

230 Edmund B. JFilson. 

lar or even amoeboid, and I have sometimes observed even trocho- 
blasts of very irregular form. But this is not the case with most 
of the isolated cells in Patella and Dentalhim, and I am inclined 
to accept the probability that the cells in question may really be 
mesenchyme- and perhaps actually muscle-cells, that have dif- 
ferentiated in more or less complete isolation from their fel- 
lows. If this be considered an improbable conclusion, it should 
be recalled that a trochoblast is probably, to say the least, as 
highly differentiated as a mesenchyme cell; yet it has been strictly 
proved that such a cell may undergo its normal differentiation 
and continue for a time to perform its normally coordinated ac- 
tivities when completely isolated from the time of its formation. 
Further research specifically directed to this point will, I believe, 
give a positive result on this very interesting question. 

5. Summary en isolated cells from the lower hemisphere. 

The evidence derived from these cells is less detailed and com- 
plete than that derived from the first quartet; but as far as it 
goes gives the same general conclusion. The isolated >^-macro- 
mere, yV-macromere or second quartet-cell segments as if form- 
ing part of a whole embryo, and shows more clearly than do the 
first quartet cells that not only the form, but also the rhythm of 
cleavage is maintained (precisely as I showed in the nemertlne) ; 
for In the cleavage of both the ^s- and the yV-macromere the 
fourth quartet cell is the last to form. Only the embryos con- 
taining derivatives of the second quartet produce secondary 
trochoblasts, namely, those arising from the ^-macromere or the 
second quartet-cell. While all the embryos close, only those gas- 
trulate that contain the basal region (/'. e., the entoblast region). 
All of the three types examined develop one or two feebly cil- 
iated cells that probably represent cells of the pre-anal ventral 
ciliated tract. Finally, there is some evidence, though only of an 
inferential character, that Isolated mesoblast-cells may develop 
into mesenchyme-cells, possibly into muscle-cells. We may, there- 
fore conclude that, speaking broadly, the development of cells 
of the lower hemisphere, like that of the upper, conforms to the 
mosaic principle. 

Experimental Studies on Germinal Localization. 231 


I have purposely left to the last an account of the development 
of the half or quarter embryos, since this is in Patella in some 
respects the least satisfactory part of the work. This is owing 
especially to the great susceptibility of the Yi and M-larvse, which 
frequently break up into smaller fragments, go to pieces, or be- 
come quite abnormal during the cleavage process, so that very 
few satisfactory larvae were obtained. In Dentalium the results 
are much better, since the blastomeres can be easily separated 
without the use of the calcium-free water; but even here my 
fixed material has proved insufficient for a satisfactory analysis 
of the internal phenomena. For these reasons the following ob- 
servations remain somewhat fragmentary and must await a sup- 
plementary study in these or other forms. 

I. Tlie partial cleavage in Dentalium. 

In my preceding paper I have described in a general way the 
development of isolated halves and quarters in Dentalium, and 
will here only add some details regarding their mode of cleavage, 
which are hardly more than a confirmation of Crampton's results 
on Ilyanassa. As in Patella, these earlier blastomeres, like the 
later ones, become perfectly spherical after isolation before cleav- 
age begins; their characteristic partial cleavage must therefore 
be due to internal factors and not to their shape. 

The AB (anterior) half, which shows only an upper white 
polar area (Fig. 83), segments equally into two, with no trace 
of a polar lobe (Figs. 84-85) , and then forms by dexiotropic divi- 
sions two slightly smaller micromeres of the first quartet, which 
are composed entirely of white material (Fig. 86). The fol- 
lowing division (Fig. 87) is like that occurring in half an egg, 
the upper cells dividing slightly unequally to form below the 
two primary trochoblasts ( la^ and ib") and above the two upper 
cells (la^ and ib^). The lower cells in the meantime produce 
the two cells of the second quartet (2a, 2b) in characteristic 


Edmund B. JVilson. 

fashion, these being likewise composed mainly or wholly of white 
material (Fig. 87). Beyond this point (16/2 stage) I have not 
followed the divisions. 

Fig. IX. 

Cleavage of Isolated Blastomercs in Dentalium. 

83, 88. isolated AB and CD halves, before division; 83-87, cleavage of AB- 
half; 84, 85, 4/2-cell stage, from, the side; 86, 8/2-cell stage, from the side; 87, 
16/2-cell stage, from the side; 88-93, cleavage of CD-half; 88-90, first cleavage 
(second polar lobe) and resulting 4/2-cell stage, from the side; 91, second 
cleavage (third polar lobe) from the side; 92, resulting 8/2-cell stage, oblique 
view, from the side and above; 93, 16/2-cell stage, with first somatoblast, 2d 
(X) obliquely from the side; 94, 2-cell stage of C-fourth; 95-99. cleavage of 
D-fourth; 95, 96, trefoils; 97, 98, 8/4-cell stages; 99, 16/4-cell stage (seen from 
the inner side, so as to appear reversed). 

Experimental Studies on Germinal Localization. 233 

The CD half, which clearly shows both upper and lower white 
polar areas (Fig. 88), forms a polar lobe from the lower white 
area and passes through a trefoil stage, nearly similar to that of 
a whole egg (Fig. 89). Measurements show, however, that the 
polar lobe is always proportionately larger than in a normal tre- 
foil, being often as large as in a whole egg, though sometimes 
more or less reduced. The lobe subsequently fuses with the pos- 
terior cell, D, producing a 4/2-cell stage closely similar to a nor- 
mal 2-cell stage, except that the inequality is greater (Fig. 90). 
At the second cleavage the polar lobe forms again (Fig. 91) 
from the larger cell, D, which divides unequally and dexiotropl- 
cally to form id, while the smaller cell, C, divides slightly un- 
equally to form ic. As in the whole egg the polar lobe then 
fuses with D, producing an 8/2-cell stage that is essentially like 
the posterior half of a normal 8-cell stage (Fig. 92). The fol- 
lowing cleavage is especially interesting, corresponding again 
with the divisions in the posterior half of a whole egg (Fig. 93). 
All the divisions are leiotropic. The two upper cells divide 
slightly unequally to form the two primary trochoblasts ( ic^, id^) 
and the slightly larger upper cells (ic^ and id^). From iC 
arises the right cell (2c) of the second quartet, while from iD 
arises the first somatoblast (2d), which Is as large as in a whole 
embryo, and in like manner is mainly formed from the lower 
white area in \D. The 16/2-cell stage has, therefore, exactly 
the same origin and composition as the posterior half of a whole 
egg, consisting of six white ectomeres (ic\ id\ ic", id", 2c and 
2d), of which 2d is the largest, and of two macromeres (2C, 
2D), which contain all of the pigment and show each an upper 
white area (Fig. 93). 

The history of isolated 54-blastomeres is entirely analogous. 
The A, B or C quadrant typically divides slightly unequally, with- 
out a polar lobe (Fig. 94), the smaller cell being composed of 
white material and the pigment remaining in the larger; but cases 
are not infrequent In which the division is nearly or quite equal. 
The D-quadrant, on the other hand, forms a polar lobe, which, 
as in a whole embryo, is typically much smaller than either the 
first or the second (Figs. 95-96). The 2-cell stage is very un- 

234 Edmund B. JFilson. 

equal, the small cell (id) being pure white, the larger showing 
both upper and lower polar areas (Figs. 97-98). At the sec- 
ond division (virtual fourth) the second somatoblast (2d) forms 
from the lower polar area, while the micromere produces the 
single trochoblast (id"), and the corresponding larger upper cell 
id^ (Fig. 99). Beyond this the cleavage was not followed in 
detail. It is noteworthy that in the divisions both of the halves 
and the fourths the normal inequality of the cells is frequently 
reduced, and this is frequently expressed by a reduction in the size 
of the polar lobe, both in the CD-half and the D-fourth — in- 
deed, I have seen only a few D-fourths In which the polar lobe 
was of full normal size, and the first division of these cells is fre- 
quently irregular and abnormal. This is doubtless due in part 
to shock, perhaps also to the effect of the calcium-free water when 
this is used. Nevertheless, I think it probable that the effect may 
also be due in part to disturbances in the arrangement of the 
cytoplasmic materials, which may possibly be interpreted as a 
regulative phenomena. 

2. The partial cleavage in Patella. 

The cleavage of isolated halves or fourths in Patella is entirely 
in agreement with the foregoing in being strictly partial in charac- 
ter, but I wish especially to emphasize the fact that, precisely as 
I showed in Cerebratulus, two general types exist, in one of which 
the cells so shift as to produce a closed embryo from the begin- 
ning, while in the other the blastula is at first widely open on one 
side. The point is important because the effect of the displace- 
ment in the closed type is to shift the primary trochoblast-groups 
more or less widely, sometimes to opposite sides of the embryo, 
while in the open type they remain in nearly their normal position. 
Nature thus performs an experiment in the displacement of the 
blastomeres closely similar to those carried out by Fischel on the 
ctenophore-egg, and a corresponding result is produced that 
clearly shows the differentiation of the cells to be independent of 
their position in the embryo. 

Experimental Studies on Germinal Localization. 235 

In the following description the typical case is described; at- 
tention is again called to the fact that the unequal divisions are 
frequently less unequal than in the normal and sometimes be- 
come quite equal. 

Fig. X. 
Isolated 1/2-BIastovicres, Patella; x 250 
100, successive camera outlines, showing relative sizes of whole egg, and 
the y2, % and ^^-'blastomeres; loi, 4/2-ceIl stage; 102, 8/2-cell stage; 103, 
16/2-cell stage, nearly typical open type, from above; 104, slightly less open 
form from the side; 105. 106, 16/2-cell stages, closed type; 107, 32/2-cell stage, 
closed type; 108, open blastula, from the open side. 

236 Edmund B. Wilson. 

The Isolated ^^-blastomere first divides equally (Fig. loi), 
then unequally and dexiotropically, so as to form two slightly 
smaller micromeres, displaced towards the left (Fig. 102). Up 
to this point the embryo remains strictly a half of the correspond- 
ing 8-cell stage. At the succeeding division the differences be- 
tween the open and closed types become apparent. In the former 
case, as shown in Figs. 103-104, the divisions may occur nearly 
typically, though frequently the cells become more or less dis- 
placed. In the second case the cells shift during the division, so 
as to fit accurately together; and, as is clearly shown in Figs. 105, 
106, the two trochoblasts (shaded) may thus come to lie on op- 
posite sides of the embryo, as is also the case with the two cells 
of the second quartet. I have not followed out in full the later 
cleavage of these larvae, which are very puzzling in both cases, 
owing to either the initial or subsequent shiftings. But so much 
is certain, that from the open type may arise an open blastula 
(Fig- 108), while the closed type remains closed; and the effects 
are clearly shown in the resulting larvae. Fig. 107 shows a closed 
32/2-cell stage, with the polar body in position. This embryo is 
diflicult to analyze in detail, but very clearly shows two rosette- 
cells above, with the corresponding i^- cells, and on each side 
are two cells that doubtless represent the daughter-trochoblasts. 
The eight cells of the lower hemisphere are more difficult to 
identify, and the cell-connections shown in the figure are only in- 

The isolated M blastomere first divides unequally, forming a 
micromere above, a macromere below (Fig. 109) ; and this is 
followed by a leiotropic division identical with that occurring 
in a single quadrant of a whole embryo (Fig. no). The 16/4- 
cell stage then divides dexiotropically, producing a 32/4-cell stage 
that may pretty accurately correspond with a single quadrant of 
the normal 32-cell stage. Fig. in shows this stage of the same 
individual shown in iio; this differs from a single quadrant of a 
whole 32-cell stage only in the fact that 2^ has extended upwards 
somewhat, so as to separate i^- from i"". Fig. 27 shows a 32/4- 
cell stage that has separated somewhat (the cells are shown ex- 
actly as they lay). Every cell is of correct proportion and po- 

Experimental Studies on Germinal Localization. 237 

sitlon, except that one of the groups of four has turned over 
(doubtless during the removal with the pipette), so that the 
upper group presents to view the outer, the lower one the inner, 
side; while the rosette-cell is somewhat too large. Both the cases 
figured represent the open type, which appears to be the rule in 
the quarter cleavage. 

3. The half and quarter larva in Patella. 

The detailed study of the larvae derived from the 3^ or K- 
blastomeres presents many practical difficulties. While the early 
cleavage of these embryos is easily determined, the later stages 
are exceedingly difficult to follow, owing to the shiftings of the 
cells, the more or less complete closure of the embryos, and the 
great number of defective or monstrous forms, I must admit that 
as far. as Patella is concerned, and in some respects in Dentalium 
also, the following account is far from satisfactory, especially in 
regard to the most interesting question of all, that of the meso- 
blast; but since I may have no opportunity to complete it at pres- 
ent, I desire to record some observations which may at least open 
the way for a more adequate study in the future. 

The most essential point has been recorded in my preceding 
paper, namely, that in Dentalium neither the >4- nor the 34-blas- 
tomere is able to produce a perfect dwarf larva; and, further, 
that the AB and the CD halves show definite and constant dif- 
ferences, the former lacking both the post-trochal region and the 
apical organ, while both these structures are present in the CD 
larva. In like manner, among the quarter larvae only the D- 
fourth produces these two structures, which are entirely lacking 
in the A, B or C-fourths. 

In Patella the corresponding comparison is far more difficult, 
owing partly to the equal size of the halves or quadrants, but 
more especially to the even greater difficulty of rearing the larvae, 
which very frequently go to pieces during the late cleavage stages, 
and invariably become irregular and monstrous during the sec- 
ond day, and finally disintegrated before the larval characters be- 
come clearly marked. My observations clearly show one point, 

238 Edmund B. JVihon. 

however, in the comparison of the two halves from the same egg, 
in which Patella differs from Dentalium, namely, that both halves 
develop an apical organ; and while this has not been directly 
proved for the four quarters, the fact, described above, that any 
isolated micromere of the first quartet may develop an apical or- 
gan leaves practically no doubt that the same is true for the quar- 
ters. The basis of the apical organ in Patella must, therefore, 
be symmetrically divided by the first two cleavages, while it re- 
mains undivided in Dentalium, remaining as a whole in the D- 
quadrant. This is possibly correlated with the fact that the apical 
rosette, formed at the fifth cleavage of Patella nnd Trochiis, fails 
to appear in Dentalium, where the i^-^ cells are as large as their 
sister-cells i^". 

In Dentalium, as described in my first paper, the larvae in- 
variably close sooner or later, and the prototroch, in most if not 
all cases, closes also to form a complete belt encircling the body. 
In Patella, however, this is not always the case; and frequently 
the >4-larv2e of 24 hours show the prototroch as an area of char- 
acteristic trochoblasts extending around one side only, terminating 
abruptly to leave a space occupied by much smaller non-ciliated 
cells. ^ (Figs. 117, 118). In other half larvae the prototroch ap- 
pears as a complete belt, in still others as a more or less irregular 
or interrupted structure. 

An examination of the earlier ciliate^l stages, combined with 
the results obtained with isolated trochoblasts, gives the obvious 
explanation of these differences. In those of the earlier larva? 
(8-10 hours) that are still open on one side (and hence must 
have been derived from the open type of cleavage) two adjoin- 
ing groups of trochoblasts are found on one side, leaving a space 
on the opposite side free from trochoblasts (Fig. 114). In the 
closed embryos, on the other hand, two corresponding trocho- 
blast groups are formed on opposite sides of the embryo, with 
only rather narrow gaps between them (Figs. 11 3- 11 5- 116). 
Both these types may be represented in twins from the same egg, 
a case which I am fortunately able to show by Figs. 113 and 114 

iln agreement with Crampton's observation that the i^-larvse of Ilyanassa 
form "a partial circle of cilia" (96, p. 9). 

Experimental Studies on Germinal Localization. 239 

(from acetic-glycerine preparations). Of these twins one (Fig. 
113) is closed and shows the gastrulation well advanced (the su- 
perficial ectoblast-cells of the lower hemisphere are only in part 
shown). This larva shows very clearly the two groups of pri- 
mary trochoblasts on opposite sides of the egg, at t and t, with 
at least two secondary trochoblasts lying between them on each 
side, the general arrangement being similar to that shown in Fig. 
116.^ The twin larva (Fig. 114) is still widely open on one 
side; and while the small ectoblast cells have closed in to fill the 
gap above, the two primary trochoblast groups lie at one side, 
leaving a wide gap occupied by smaller cells. Fig. 115 is a J-^- 
larva, which, though somewhat asymmetrical, is clearly of the 
closed type (the superficial post-trochal ectoblast-cells are shown 
only in optical sections at the sides) ; and here, too, the primary 
trochoblast groups lie on opposite sides of the larva. 

It is hardly possible to doubt that these two types of larvae 
arise from the open and closed types of cleavage, the trochoblasts 
having undergone their normal differentiation whether displaced 
or not. This has not been strictly proved by isolation experi- 
ments ; but in view of the demonstrated fact that the trochoblasts 
differentiate typically If wholly separated from their fellows, 
there can be no doubt, I think, of the interpretation offered. It 
is quite clear that in this case the prospective value of the cell is 
not a function of its position, but is dependent on its internal or- 
ganization irrespective of its position. This result is exactly anal- 
ogous to those obtained by Fischel ('98) by displacing the micro- 
meres of the ctenophore egg — an operation that, as he shows in 
the most convincing manner, leads to a correspoiiding displace- 
ment of the rows of swimming plates in the larva. 

With the M-larvas in Patella I had little success, since they al- 
most invariably broke apart into smaller fragments. A very few 
nearly complete larvae of 24 hours were, however, obtained, one 
of which is shown in Fig. 112. This larva shows a central mass 
of rather large rounded cells completely surrounded by ectoblast, 
and has evidently gastrulated. At one side is a very distinct group 

iThe latter larva apparently shows five primary trochoblasts on one side 
— a fact for which I cannot account. 


Edmund B. Ifllson. 

Fig. XI. 

Experimental Studies on Germinal Localization. 241 

1/4 and 1/2 Larva. Patella; x 250. 

109-111, cleavage of isolated 1/4, open type; 112, resulting larva, 24 hours; 
113, 114, tw^in larvae. 9 hours, 113 of the closed type, 114 of the open; 115, 
closed J^-larva, 9 hours; 116, closed J/^-larva, 11 hours, apical view; 117, 118, 
products of open type, 24 hours. 

242 Edmund B. Wilson. 

of six trochoblasts. Four of these, bearing powerful rows of 
cilia, are evidently primary trochoblasts; two, lying in front of the 
last, are much smaller, and are probably secondary trochoblasts 
from the first quartet. Those of the second quartet have either 
failed to develop or have broken away from their connections (as 
very often occurs with all the trochoblasts owing to their ac- 
tivity). No apical organ was seen in this larva; but I observed 
an apical organ in several less normally developed individuals, 
and since the apical organ constantly appears in the J/g-micromere 
larva, there can be no doubt that it may appear also in the J4- 

Perhaps the most interesting question presented by these larvas 
is whether the AB and the CD half-larvae differ in respect to the 
mesoblast; for if the mosaic principle holds for this structure, 
one should expect to find coelomesoblast only in the CD half. 
For the present I can give no certain answer to this question, fur- 
ther than to state that in DentaJium the two larvae certainly dif- 
fer to some extent in respect to the mesoblast, and there is possibly 
some reason to conclude that they do also in Patella. 

In the latter form some of the larvae show a large rounded cell 
in the upper region of the central mass (the dotted outline in 
Figs. 113 and 115) which does not appear in others; and this 
difference distinctly appears between the two twin larvae shown 
in Figs. 113 and 114. This cell is possibly the primary meso- 
blast, 4d; but it may also represent the large rounded cell which 
I have considered to be 4D In the normal larva (Figs. 15-16). 
This evidence, unsatisfactory as It is. Is mentioned as an Indica- 
tion that the Internal structure of the two half-larvae shows dis- 
tinct differences In Patella. In Dentalium the evidence is some- 
what better, but still far from adequate, owing to paucity of ma- 
terial and the confused appearance of the inner cell-mass, as seen 
either In total preparation or In sections. Sections of the CD 
larvae nevertheless show groups of smaller and irregular cells 
lying between the large entoblast-cells and the ectoblast, and 
there is little doubt that these represent in part the coelomeso- 
blast. Sections of the AB larvae are in general closely similar to 
those of the lobeless larvae described in my preceding paper, but 

Experimental Studies on Germinal Localization. 243 

in some cases distinctly show a few small cells lying between the 
gut and the ectoblast. The only conclusion that I am justified in 
drawing is that the mesoblast cells are more numerous in the CD 
larva than in the AB, and additional material will be necessary 
to determine the point. When, however, we consider the evi- 
dence, not entirely conclusive but still fairly definite, given in my 
preceding paper, that the material of the polar lobe (which 
passes only into the CD half) is necessary for the production of 
the coelomesoblast, I think It may be concluded with some prob- 
ability that the mesoblast-cells (If they be such) found in the AB 
half represent a portion of the larval mesoblast or ectomesoblast, 
and that the coelomesoblast Is represented only in the CD half. 
I hope in the near future to obtain additional material that may 
afford a more definite conclusion. 


{This Applies Primarily to Patella.) 

1. Isolated blastomeres of any stage segment essentially in 
the same manner as if still forming part of a complete embryo, 
with a tendency, however, for all unequal divisions to be less un- 
equal than In the normal. The partial form of cleavage is fre- 
quently masked by shifting of the cells. 

2. All of the partial embryos, if of sufficient size, tend to close 
to form morula- or blastula-like structures ; but these only gastru- 
late if they contain entoblast material from the basal cells. Apart 
from such closure all of the cells, and their products, as far as 
examined, differentiate typically, regardless of their relative po- 
sition or of complete Isolation from their fellows. 

3. Isolated >^-mIcromeres produce pyriform larvs, bearing 
at one end an apical organ, at the other a group of four primary 
and two secondary trochoblasts. In Dentalium the apical organ 
is produced only by the posterior micromere, id. 

4. Isolated primary trochoblasts (i') divide twice and pro- 
duce four typical ciliated prototrochal cells. Isolated first prod- 
ucts of the primary trochoblasts (i"S i"'") divide once and pro- 

244 Ed?fiiind B. inisoH. 

duce a pair of typical prototrochal cells. Isolated second prod- 
ucts (i"-^"S i"'^", etc.) undergo no further division, but different- 
iate singly into typical prototrochal cells. 

5. Isolated i^ cells produce embryos bearing at the anterior 
end an apical organ, at the other two secondary trochoblasts. 

6. Isolated products of the i^ cells differentiate into typical 
sensory cells of the apical organ, into secondary trochoblasts, and 
into less differentiated ectoblast cells. 

7. Isolated >^-macromeres produce closed embryos that gas- 
trulate and bear at one end one or two secondary trochoblasts, 
and at some other point a small group of feebly ciliated cells, 
probably representing the pre-anal ciliated cells of the normal 

8. Isolated yV-macromeres produce closed embryos that 
gastrulate, bear no trochoblasts, but have feebly ciliated cells, as 
in 7. 

9. Isolated cells of the second quartet produce closed ecto- 
blastic embryos bearing one or two secondary trochoblasts, and 
one or two feebly ciliated cells, that probably also represent part 
of the pre-anal tract. These embryos do not gastrulate, but may 
form mesenchyme-like cells. 

10. Isolated >4-blastomeres produce embryos that gastrulate, 
produce four primary trochoblasts, at least two secondary ones, 
and an apical organ. 

11. Isolated ^-blastomeres produce, in Patella, larvae bear- 
ing an apical organ, and a prototroch, either open or closed, ac- 
cording to the mode of cleavage. In DentaUiim only the CD 
half produces an apical organ and a post-trochal region, and 
probably only this half produces caelomesoblast. 

12. The development of both Patella and Dentaliuvi is essen- 
tially a mosaic-work of self-differentiating cells. 


The experimental results brought forward in this paper and the 
preceding one seem to me to establish definitely the principle of 
mosaic development in the case of the mollusks Dentalium and Pa- 

Experimental Studies on Germinal Localization. 245 

tella, and to place the study of cell-lineage on a new and firmer 
basis. Clearly as the exquisite adjustment between the cleavage- 
process and the operations of morphogenesis has been revealed 
by the descriptive-comparative study of cell-lineage, it appears in 
still stronger relief in the light of the experimental proof that 
the cleavage-pattern, as a whole and in detail, is the visible ex- 
presson of an actual distribution of specific morphogenic factors 
among the cells. 

Although Crampton's initial, and hitherto almost unique, ex- 
periments on this type of development had led' to the expectation 
that some evidence of cell-specification and self-differentiation 
would be found, I confess that I was not prepared to find that evi- 
dence so circumstantial and consistent. The evidence in Patella 
that the cleavage-cells are definitely specified from the time of 
their first formation, and. that they undergo self-differentiation 
without essential modification through their relation to the other 
cells, is demonstrative in the case of the cells of the first quartet, at 
least as far as the i6-cell stage, as shown by the development of 
isolated entire micromeres at the 8-cell stage, and of their prod- 
ucts i^ and i^ at the i6-cell stage. It is no less demonstrative in 
the case of the products of the primary trochoblasts isolated at 
the 32- and 64-cell stages; and inasmuch as cells of the apical or- 
gan derived from the i^-^ cells, and secondary trochoblasts derived 
from the i^- cells, also differentiate typically when the isolated 
micromere is allowed to segment continuously in the calcium- 
free water, and the cells are separated more or less completely 
after every division, the conclusion is unavoidable that these cells, 
too, may undergo their characteristic development in complete 
isolation from their fellows. Less detailed, but hardly less con- 
vincing, is the evidence derived from the isolated ]4. basal, the 
tV basal, or the isolated second quartet-cell ; and it can hardly 
be doubted that the individual products of these respective cells 
are, like those of the first quartet, definitely specified in greater 
or less degree. 

The general conclusion thus reached in the case of Patella is 
sustained by the development of larger masses of cells derived 
from the earlier stages both of Patella and of Dentaliiim. The 

246 Edmund B. JFilson. 

entire first quartet of Patella, when isolated, produces a mass of 
ectoblast-cells, which, though It closes, does not gastrulate, but 
undergoes essentially the same differentiation as if It formed the 
upper hemisphere of a complete larva. The isolated quadrant 
of a 4-cell stage gastrulates, produces a group of trochoblasts and 
an apical organ, the latter structure appearing apparently in any 
of the quadrants in Patella, while in Dentaliuin It Is restricted to 
the D-quadrant. In Dentaliiim, further, only the D-quadrant 
produces a post-trochal region, which Is due to the fact that this 
quadrant alone contains the material of the lower polar area from 
which arises the somatoblasts. Finally, the two halves of the 2- 
cell stage gastrulate, but (at least in Dentalhim) differ widely in 
their later development. Both in Dentalium and in Patella the 
half-embryo forms a prototroch, which In the former seems al- 
ways to close to form a complete ring, but In Patella frequently 
remains open at one side, forming a half ring. In Patella both 
halves form an apical organ; in Dentalium only the CD-half. 
In Dentalium, finally, only the CD-half forms a post-trochal 
region, for the same reason as in case of the D-quadrant. It is 
probable, further, that only the CD-half and the D-quadrant pro- 
duce coelomesoblast. This conclusion has not thus far been sat- 
isfactorily established by direct examination of the half-embryos, 
but is Indirectly rendered very probable through the observations 
on the lobeless larvae recorded in my preceding paper. 

The foregoing facts constitute a strong body of prima facie 
evidence that the entire cleavage-pattern in the molluscan egg 
represents (with certain reservations considered beyond) a mosaic- 
work of self-differentiating cells, exactly in the sense of Roux's 
general conception.^ The proof is indeed entirely complete In 

^Here, and in all that follows, I exclude from that conception the hypothe- 
sis of qualitative nuclear division. It should be borne in mind that Roux 
himself expressly stated as early as 1893 that this hypothesis did not form a 
necessary part of his conception. "Die beiden AmuiJimefi" (nuclear idioplasm, 
distribution by qualitative division) "sind jedoch nicht unerlasslich noth- 
wendige Glieder mciner in iliren wesentlicJten Theilen experimentell erwles- 
enen Auifassung;" ('93.2, p. 874, Italics in the original). This fact has been 
ignored by many of Roux's critics, in spite of the fact that some of his most 
important contributions to experimental embryology have been specifically 

Experiniental Studies on Germinal Localization. 247 

the case of only a few kinds of cells. It is evident that the limi- 
tations of potency vary in different cells — the i^ cells, for exam- 
ple, contain more complex potencies than the 1= — and It is quite 
possible that dependent or correlative differentiation may play a 
larger part in the development than my experiments have thus 
far shown. But the proof by experiment of definite specification 
and self-differentiation in only a few categories of the early cleav- 
age-cells establishes a principle that is to be reckoned with as a 
most important factor in the whole problem of embryonic differ- 
entiation. If, in considering some aspects of this problem, I 
again take up a discussion that has been so prolonged, it is because 
I believe that the importance of the principle of mosaic-develop- 
ment, and of the nearly related one of specific formative or deter- 
mining stuffs, has received insufficient recognition by many embry- 
ologists, and by some has been prematurely discredited. That a 
reaction is well under way will be evident to everv reader of Fis- 
chel's ('03) excellent recent discussion of development and dif- 
ferentiation, the essential conclusions of which agree closely with 
those earlier stated in a brief form in my paper on cleavage and 
mosaic-work ('96, appended to Crampton's paper), and more 
fully considered in the discussion of my nemertine results ('"03) ; 
the agreement with the conclusion reached in this and the preced- 
ing paper is still closer. In full harmony with the same general 
conception are the important cytological results of Lillie ('99, 
'01) and Conklin ('98, '99, '02) cited in the two preceding 

The long continued discussion of the mosaic-theory of develop- 
ment that followed its first definite formulation by Roux in 1888, 
in the course of which Roux so ably defended his position, has 
been greatly prejudiced by the fact that the experimental analysis 
of cleavage was at first confined to the so-called "indeterminate" 
types of cleavage, such as those of the echinoderm, the medusa 
and Amphioxus ( it may for the time be left an open question 
whether that of the frog should be placed in the same class) . The 

directed towards the role played in development by the segregation and locali- 
zation of cytoplasmic materials. Roux himself ('03) has now abandoned the 
second of these "Annahmen" (qualitative nuclear division). 

248 Edmund B. PFilson. 

earlier results formulated for these types seemed wholly subver- 
sive of the mosaic-principle and of the nearly related one of ger- 
minal prelocalization. In the sea-urchin egg, the first in which 
an isolated blastomere was shown to be capable of producing a 
complete dwarf larva, the experiments seemed at first to show 
that the blastomeres are composed of indifferent material, so that, 
to cite an early statement of Driesch's, "Durch die Theilung bei 
der Furchung vollig gleichwerthige, zu allem fahige (indiffer- 
ente) Stiicke geschaffen werden" ('92.2, p. 36), forming a ma- 
terial "welches man in beliebiger Weise, wie einen Haufen Kug- 
eln durch einander werfen kann, ohne dass seine normale Ent- 
wicklungsfahigkeit darunter im mindesten leidet" {op. cit. p. 
25). Despite the fact that Driesch early recognized that the 
cytoplasm of this egg is not isotropic, he considered that his ex- 
periments definitely overthrew His's principle of "Organbildende 
Keimbezirke," or at least deprived it of all casual significance in 
the echinoderm egg ('92.1, p. 178, '93, p. 243). Specification of 
the early cleavage cells was denied ('92.2, p. 22), as was also the 
principle of mosaic development as applied to this egg (1. c, p. 
36). Again, in the paper of 1899, on "Die Localization mor- 
phogenetischer Vorgange," where Driesch's theory of vitalism is 
first definitely formulated, the ground is taken that "Darin nam- 
lich, das jeder beliebige Eitheil, sowie das Eiganze in beliebiger 
Verlagerung eine ganze Larve llefern, also jede 'Organization,' 
die postulierte Vorbedingung zum Eintritt lokalisirten speci- 
fischen Geschehens iiberhaupt, nach Storung, regulatorisch wieder 
herzustellen vermag, kommt zum Ausdruck, das eben die 'Struck- 
tur' des Eies nicht aus mannigfach-verschiedenen Elementen in 
irgendwie typisch-specifischer Lagerung aufgebaut sein konne, die 
etwa zu den spateren Differenzierungen in irgend einer Bezie- 
hung stiinde" {op. cit., p. 43). It is hardly necessary to point 
out how greatly all this has been changed by Boveri's discovery 
of the fact that the sea-urchin egg does in point of fact contain 
"mannigfach-verschiedene Elemente" disposed in a "typisch- 
specifischer Lagerung," which are proved by the experiments of 
both Boveri and Driesch to stand in definite relation to the sub- 
sequent process of cleavage and differentiation. The relation of 

Experimental Studies on Germinal Localization. 249 

these facts to those determined in the nemertine and mollusk is 
considered beyond. I will at this point only express my agree- 
ment with the conclusion of Fischel, that "Sowohl bei den Echin- 
odermen, wie bei den speciell so-genannten 'Mosaik-eier' erfolgt 
die normale Entwickelung im Wesentlichen als Mosaik-arbeit" 
(Fischel, '03, p. 728). I am convinced that had the experimental 
analysis of cleavage been first undertaken in the case of such a 
determinate type as that of the gasteropod or annelid, and had 
Roux not handicapped his theory with a purely speculative hy- 
pothesis of differentiation, which proved to be untenable, the 
whole discussion would have taken a very different course ; and I 
believe it would from the first have been recognized that the 
mosaic-principle holds true in greater or less degree for every type 
of development, not excepting the most "indeterminate" forms 
of cleavage. 

My experiments on the unsegmented egg of Dentalinm have 
added fresh proof to that obtained by Fischel and his predeces- 
sors in the ctenophore, that the cleavage-mosaic is a mosaic of 
specifically different cytoplasmic materials, in which are somehow 
involved corresponding morphogenic factors. In this egg, con- 
firming and extending the earlier work of Crampton on the gas- 
teropod-egg, I was able to show even more definitely than has been 
done in the ctenophore-egg the existence in the unsegmented egg 
of prelocalized cytoplasmic regions, distinguishable by the eye, 
that stand in some necessary relation to the formation of the 
structures to which they give rise in the normal development; for 
if one of these areas (the lower polar area) be removed, the 
structures to which it is destined to give rise fail to develop, while 
if this area remains while other areas are removed the structures 
in question make their appearance. His's principle of "Organ- 
bildende Keimbezirke," which he developed in a purely descrip- 
tive sense, is thus shown to have a true causal significance. Since, 
further, this area contains no nucleus, the conclusion is unavoid- 
able that here, as in the ctenophore — and as we are now able to 
say, even in the echinoderm — there is a localized distribution to 
some extent, of the factors both of cleavage and of differentia- 
tion in the cytoplasm before development begins. A no less sig- 

250 Edmund B. Wilson. 

nificant fact, proved by these experiments in connection with ob- 
servations by many observers of normal cell-lineage, is that the 
germ regions prelocalized in the unsegmented egg are, at least 
in the case of certain cells, accurately marked off by the subsequent 
lines of cleavage. This is shown with great clearness by the his- 
tory of the lower polar area in DentaUum (or the analogous 
lower green area in Myzostoma or the lower polar ring in Rhyn- 
chelmis), which, although it lies primarily at the center of the 
lower hemisphere, is not bisected by the first or the second vertical 
cleavage (despite the fact that both the cleavage furrows first lie 
exactly in the egg-axis), but is moved to one side so as to pass 
bodily into one of the cells at each division. Here is an adjust- 
ment, of admirable accuracy, by which a specific prelocalized area 
is handed on from cell to cell, to be finally assigned to its proper 
positon in the cell-mosaic; and if such be the case with one such 
specific germ area, we have strong ground to infer that it is also 
so with others. In such cases as these it is evident that the fac- 
tors of cleavage run so accurately parallel to those of differentia- 
tion that they must be referred to a common determining cause, 
and may be treated as practically identical. 

But even in cases where the adjustment is less evident, or less 
precise (as appears to be the case, for example, in the third cleav- 
age of the echinoderm egg, considered beyond) we shall not, I 
believe, escape the conclusion that cleavage involves a definite 
distribution of specific morphogenic factors among the cleavage 
cells. The facts, proved by my experiments, that these factors 
may be completely separated and isolated by cell-division, and 
may retain their specific character after isolation of the cells, are 
only intelligible under the assumption that they are somehow in- 
volved in specific materials or stuffs which differ in a definite way 
and have a specific topographical grouping in the undivided egg. 
This conclusion is not to be avoided by assuming that the visible 
cytoplasmic differences are only an accompaniment or consequence 
of an invisible ulterior structure or organization. Admitting this, 
and even admitting, for the sake of argument, that the localized 
cytoplasmic factors are not definitely characterized chemical ma- 
terials, but only local physical or structural conditions, established 

Experimental Studies on Germinal Localization. 251 

by virtue of the relation of the particular cytoplasmic regions to 
the egg as a whole : the fact remains that the cytoplasmic sub- 
stance possesses different specific qualities in different regions, and 
that these differences persist after the regions have lost their re- 
lation to the whole. Only by a play upon words, therefore, can 
the conclusion be escaped that the cytoplasmic regions consist of 
specifically different substances having a definite morphogenic 
v^alue. The question whether these substances are to be consid- 
ered as preformed building materials, or rather as specific deter- 
mining materials^ (such, for example, as enzymes) is a second- 
ary one, on which I do not propose to enter here. Holding both 
these possibilities in view, I can see no valid objection to the frank 
adoption, in a provisional sense, of the term "formative stuffs" 
in the general spirit of the Bonnet-Sachs hypothesis, awaiting fu- 
ture research, to determine what is their mode of action. We 
must, therefore, conclude that the cleavage-pattern represents lit- 
erally a mosaic-work of such formative stuffs that have been dis- 
tributed by the cleavage process, and that the specification of the 
cells is within certain limits determined by their inclusion of these 
stuffs. If for the conceptions of qualitative and quantitative 
nuclear divisions we substitute those of qualitative and quantitative 
cytoplasmic divisions, a very large part of the development that 
Roux has given to his theory in his long controversy with Driesch, 
O. Hertwig and other writers, is, I believe, entirely valid. I 
shall not undertake to go over the whole of this ground again, 
but will apply these terms to a specific interpretation of certain 

In my preceding paper I have suggested that the difference in 
behavior between isolated blastomeres of different forms is pri- 
marily due to differences in the initial form and degree of segre- 
gation. The possibility of the production of a perfect larva from 
either half of any quarter in the egg of Amphioxus, Echinus or 
Cerebratulus is given by the symmetrical or purely quantitative 
distribution of materials by the first and second cleavages.' In 
Dentalium both halves are not able to produce perfect larvae, 

'^Cf. Morgan, Regeneration, p. 89. 
Kf. Fischel, 03, p. 7^2>- 

252 Edmund B. JFUsoti. 

owing to an asymmetrical distribution of material, the cleavage 
being visibly qualitative from the beginning; and it is impor- 
tant to note that this asymmetry of distribution is effected by the 
process of cleavage itself, since the primary segregation-pattern, 
as far as can be determined, is symmetrical with respect to the 
axis. In the nemertine or sea urchin the first qualitative division 
occurs at the third cleavage (which is also qualitative in the mol- 
lusk)^ which for the first time separates ectoblast-stuff from en- 
toblast-stuff. A comparison of the different forms indicates, how- 
ever, that in respect to this cleavage they differ somewhat in de- 
gree. In Patella the cells of the first quartet are from the first 
completely specified, whether as a group or individually, and pro- 
duce purely ectoblastic embryos that never show any tendency to 
gastrulate. The same is true in Cerebratiilus, according to the 
recent work of Zeleny ('04), which shows that if in the 8-cell 
stage the upper and lower quartets be separated along the line 
of the third cleavage, both quartets develop into closed swimming 
embryos, but the upper one (although the larger) does not gas- 
trulate, though it produces an apical organ, while the lower one 
gastrulates but produces no apical organ." In the; sea urchin, 
however, a small proportion of the upper cells (20-25%) are able 
to gastrulate (Driesch, '00. '02.2) ; and this can only mean that 
the third cleavage is less strictly qualitative or not invariably so. 

lAs was also assumed by Samassa in the case of the frog's egg, "Diese ver- 
schiedenen Entwickelungsbedingungen konnen aber nur in verschieden Sub- 
stanzen Hegen, die bei der qualitativ ungleiche Theilung der dritten Furchung 
den beiden Zellarten zufallen" ('96, p. 386). 

2In the light of Zeleny's observations on the 8-cell stage, and in spite 
of his apparent confirmation of my own preceding results on the blastula 
stage, it seems to me very probable that the gastrulas I obtained from upper 
fragments of blastulas in Ccrebratuhis were obtained by slightly oblique sec- 
tion, so that a small group of entoblast cells were included in the upper frag- 
ment. I have since observed that the entoblast-plate extends nearly to the 
equator of the egg, so that even a slight obliquity in the plane of section 
might give a misleading result. A similar interpretation not improbably may 
apply to the upper fragments of echinoderm blastulae, cut in two just before 
gastrulation, which were observed by Driesch ('95) to gastrulate; but these 
were cut eii masse without individual orientation, and the experiments evi- 
dently do not exclude the possibilty that the upper fragments may have con- 
tained a part of the entoblast-region. A repetition of this work on both forms 
by means of individual operation is much to be desired. 

Experimental Studies on Germinal Localization. 253 

Boverl ('oi.i) has In fact shown experimentally that the ability 
to gastrulate depends on the presence of a certain amount of the 
pigment-band that approximately coincides with the entoblast- 
zone; and the variation in this regard is explicable under the as- 
sumption either of a varying position of the third cleavage-plane 
with respect to the entoblast-zone or of a variation in the de- 
gree of concentration of the entoblast-stuff. While Boveri adopts 
provisionally the former of these alternatives, he also suggests 
that the formation of entoblast and mesenchyme is not absolutely 
predetermined in the plasma, but occurs at the "most vegetative" 
point, which is the lower pole. Driesch ('02.2) adds the sug- 
gestion that the frequent failure of the animal larvae to gastrulate 
may be due, not to absolute lack of "vegetativlty" ("Um einen 
nicht sehr schonen aber deutllchen Ausdruck zu gebrauchen") , 
but to its Insufficient degree; and he has recently shown ('03) by 
an experiment of admirable ingenuity that artificial displacement 
of the third cleavage-furrow towards the vegetative pole causes 
a large Increase in the proportion of gastrulas produced by the 
isolated upper cells. This interpretation becomes perfectly intel- 
ligible If stated frankly in terms of the formative stuff hypothesis; 
and it harmonizes with my conclusion regarding the Dentalium 
egg that the influence of the specific stuffs is within certain limits 
qualitative rather than quantitative, which was based on the fact 
that If the upper part of the egg be cut away, leaving the whole 
of the lower pole area the polar lobe typically is reduced to the 
correct proportional volume, and the resulting larva has a post- 
trochal region of the proper size. This conclusion Is in agree- 
ment with that of both Boverl and Driesch, that the plasma struc- 
ture plays "nur eine determlnirende, kelne fixierende Rolle" 
(Driesch, '02.2, p. 522). 

In the ctenophore It appears from Fischel's observations ('98) 
that the first qualitative division is the fourth, which first sep- 
arates ectoblastic micromere material from the entoblast con- 
taining basal cells. In the whole series up to this point we have 
been considering a segregation that in its initial form is vertical 
and symmetrical about the axis, though in the moUusk and an- 
nelid it becomes asymmetrical in the course of the first cleavage 

254 Edmund B. Jf'Uson. 

(prov^ed by direct observation in Clepsine, Rhynchelmis, Myzos- 
toma and DentaUiim) . In the medusae it appears from the work, 
of Zoja ('95) and Maas ('01), that the primary segregation is 
not visibly polarized, but concentric; and qualitative division (de- 
lamlnation) does not take place before the fifth cleavage.^ It 
seems evident that in these differences of form and degree of the 
initial segregation pattern we find the leading principle for an 
explanation of the differences in mode of development shown by 
isolated blastomeres of the various forms ; though as pointed out 
beyond, a complete explanation Is not given by these facts alone. 
To consider one or two more detailed Instances, the first division 
of the first quartet cells in Patella Is qualitative, not merely In a 
descriptive or prospective sense, but actually, as Is proved by ex- 
periment. By the same standard, the second division of these 
cells Is qualitative In the upper cell (i^), but only quantitative 
in the lower one (i"). Such facts as these give the strongest 
ground for the conclusion that all the divisions that would be 
considered as qualitative or quantitative from thfc point of view 
of descriptive cell-lineage, are really such as regards the inherent 
factors of differentiation. The descriptive and comparative study 
of cell-lineage represents something more, therefore, than a mere 
enumeration of successive cell divisions and their geometrical re- 
lations, and has the value of a direct examination of the normal 
morphogenic process. 

These conclusions may appear to conflict with the view that 
has been frequently urged by embryologists In late years that the 
organism develops essentially as a whole, and that cell-formation 
plays but a subordinate part in the morphogenic process. The 
conflict is, I believe, only a seeming one. Roux has repeatedly 
pointed out that the mosaic-principle Is by no means Irreconcll- 

'^Cf. Maas: "Wenn die (cytoplasmic) Substanzen in alien Radien, resp. 
Axen, gleichmassig verteilt sind, wie bei den kugeligen Eiern von Medusen. 
dann und nur dann hat man ein wirklich isotropcs Ei; in anderen Fallen, wo ein 
polare Anordnung festgestellt werden kann, wie bei den echinodermen. bes- 
teht die Isotropic niir urn eine hcstimmtc Axe; in weiteren Fallen kommt durch 
Gestalt des Eies, wie bei den Cephalopoden, oder durch Lagerung der Substan- 
zen wie bei den Ampbibien, eine bilateral-symmetrisch Anordnung zu Stande 
und in anderen Fallen ist diese Anordnung noch etwas komplizierter (siehe 
Z. B. Mycosotoina)." ('03, p. 72.) 

Experimental Studies on Germinal Localization. 255 

able with such a view, and he has steadily maintained the position 
that the development of every animal presents a combination of 
self-differentiation and correlative or dependent differentiation, 
the relation between which varies more or less widely in different 
cases/ Only the most thorough experimental study can deter- 
mine what this relation is in any individual case. The hypothesis 
of qualitative nuclear division is no doubt responsible for the dis- 
favor with which the conception of self-differentiation was re- 
ceived by many writers, who either relegated it to a position of 
quite minor importance or rejected it in toto, adopting only hy- 
potheses of correlative differentiation, or advocating a less clearJv 
defined conception of the "organism as a whole," to which the 
differentiation of the cells was assumed to be subject. O. Hert- 
wig's theory of cellular interaction is a clearly formulated conr 
ception of this type, cleavage being assumed to be merely a multi- 
plicative process, producing qualitatively equivalent blastomeres 
that differentiate by cellular interaction {e. g., '92, p. 481, '93, p. 
793). "Die Zellen determiniren sich zu ihrer spateren F^igenart 
nicht selbst, sondern werden nach Gesetzen, die sich aus dem Zu- 
sammenwirken aller Zellen auf den jeweiligen Entwicklungsstu- 
fen des Gesammtorganismus ergehen, determinirt" ('98, p. 144). 
Cell-lineage, therefore, has only an incidental significance, arising 
from the continuity of development, which involves the deriva- 
tion of each part from an earlier group of cells, itself in turn the 
product of a still earlier one ('92, p. 479). Whitman ('93), in 
his singularly thoughtful and suggestive essay on the "Inadequacy 
of the Cell Theory of Development," while repudiating the 
theory of cellular interaction as such, urged with great force the 
subordination of the Individual cells in development to the or- 
ganization of the embryo as a whole — a conception which, though 
differing widely in its form of expression, has, I think, much in 
common with Driesch's theory. The same general view is very 
specifically interpreted in Child's valuable descriptive paper on 
the cell-lineage of Arenicola ('00), in such statements as the fol- 

^C/^. Roux, '88, p. 455, and elsewhere. Heider, in his suggestive survey 
of the determination-problem ("00), and Fischel, in his more recent discussion 
(03), takes the same ground. See also Korschelt and Heider, '02. 

256 Edmund B. JVilson. 

lowing: "The differences between the quiescent trochoblasts and 
the other cells does not necessarily signify that the former con- 
tain a special substance which makes them distinctively trocho- 
blasts from the time of their formation. Of course, at some time 
they do become distinctly trochoblasts, but simply because of their 
relation to the whole" (p. 664). I have cited this particular 
case, since It Is precisely In the case of the trochoblasts that experi- 
ment most Indubitably demonstrates self-differentiation Inde- 
pendently of the position of the cell In the embryo. To 
cite a more general statement, "The material separated 
as the result of precocious segregation may, I believe, be 
perfectly Indifferent material except as regards position" (p. 
682). "Certain amounts, rather than certain kinds of material, 
are stored up in certain cells just where they will be in position 
to produce by coordinated action the 'desired result' " (p. 679. 
Italics mine). I must own to some difficulty In grasping the con- 
ception of a "precocious segregation of perfectly Indifferent ma- 
terial"; but, this aside, it is clear that differentiation Is considered 
to be effected, not through the specific and Inherent nature of the 
substance of the individual cell, but through correlative action, 
the hypothesis even being advanced that an important function 
of the spiral type of cleavage Is to provide for this purpose the 
most direct and intimate possible communication between the 
blastomeres (p. 658, etc.). 

Llllle, who has contributed such valuable observations on the 
progressive segregation and organization of the egg-substance, 
and has recognized In the fullest degree the complexity of that 
organization and the Importance of precocious segregation, never- 
theless casts considerable doubt on the conception of prelocalized 
germ areas ('01, p. 269), and feels constrained to take the po- 
sition "That the entire organism in every stage of its develop- 
ment exercises a formative influence on all its parts, appears to 
me an absolutely necessary hypothesis" ('01, p. 273). I do not 
doubt, as will appear beyond, that this position, with proper quali- 
fication, is well grounded; but do not the phenomena of self-dif- 
ferentiation, as shown in the Independence of grafts or in the 
typical differentiation of Isolated blastomeres in Patella, show that 

Experimental Studies on Germinal Localization. 257 

as thus stated the conclusion is somewhat misleading? I cannot 
think otherwise. The fundamental conception that the develop- 
ment of every part is conditioned by that of the organism as a 
whole is one that every embryologist must accept; but it seems 
to me that Driesch, whom no one will consider a partisan of the 
mosaic-theory, expresses the truth when he says (Analytische 
Theorie, p. 94), "In diesem Sinne ist nun Selbst-differenzieruno- 
einmal angelegter Teile ein wesentliches Merkmal der Ontoge- 
nese; ja sie ist in Hinsicht auf die spatere Einheitlichkiet und das 
physiologische Zusammenwirken unaphangig entwickelter Gebilde 
von einem ganz eigenartigen Interesse" (Italics original). The 
fact must be recognized that the developmental energies and poten- 
cies undergo a secondary distribution among the cells or tissues 
at an earlier or later period, and in varying degrees, which in- 
volves corresponding limitations in the secondary centers thus 
created. We have long been familiar with such limitations in the 
case of the "germ-layers," though the experimental evidence has 
shown that they are here less rigid than was formerly supposed. 
They have been experimentally shown with great clearness by 
Driesch ('95) in the structures of the blastula, gastrula, and 
young larva of the echinoderm at successive stages. The ex- 
perimental results demonstrating the mosaic-character of cleav- 
age have merely shown that similar restrictions of potency may 
occur still earlier, so as to become manifest even in the early 
cleavage-cells. Now, it is clear that the primary localization of 
formative stuffs in the unsegmented egg is essentially an act of 
the "organism as a whole;" and even though a complete preform- 
ation and prelocalization of specific stuffs for every cell and tissue 
were assumed — and I believe with Boveri and Fischel that such 
an assumption is not necessary or even probable — we should not 
escape the necessity for assuming such action of the whole. That 
the egg undergoes a definite development during its ovarian his- 
tory and in the stages preceding cleavage, we have evidence both 
cytological and experimental. The character of the primary seg- 
regation-pattern thus determined is indeed determined by the egg 
as a whole, and the localization thus initiated forms the primary 
basis of the specification of the blastomeres and organs that de- 

258 Edmund B. JVihon. 

velop from the various egg regions. This is quite in harmony 
with Whitman's contention that "organization precedes cell-form- 
ation and regulates it" {op. dt.^ p. 115). But, while in agree- 
ment with the general spirit of his conclusions, as I understand 
them, it seems to me that Whitman's statement does not sufficiently 
recognize, first, the fact that the differentiating factors may un- 
dergo so accurate and complete a distribution among the cells, 
and be so largely emancipated from the general control as is 
proved by my experiments — in other words, sufficient weight is 
not given to the effects of precocious segregation; second, (and 
here I should more distinctly take issue with him) that the cyto- 
plasmic segregation or "organization" is a progressive or epigen- 
etic process. 

As regards this second point, in my preceding paper I have en- 
deavored to show that the Dentaliiim egg presents a form of pre- 
cocious segregation (and localization) which in other forms, such 
as the eggs of certain annelids, is acquired at a later period. The 
facts observed by Boverl on the Strongylocentrotns egg, and the 
experimental results of Yatsu, Zeleny and myself on Cerehratu- 
lus clearly indicate that in these forms, too, the cytoplasmic seg- 
regation is gradually effected, and at the time of the third cleav- 
age has progressed further In the nemertine than in theechlnoderm. 
There is, therefore, a legitimate basis for the conclusion that the 
degree as well as the form of segregation existing at the begin- 
ning of cleavage may vary more or less widely; and hence for the 
further assumption that the mosaic character of the early cleav- 
age stages may be expressed in different degrees. For this rea- 
son, in so far as the term "organization" as used by Whitman is 
applied to the cytoplasmic conditions, I am unable to accept his 
conclusion that the eggs of different forms do not differ In degree 
of "organization;" or that "Cell-orientation may enable us to 
Infer organization, but to regard It as a measure of organization 
Is a serious error" {op. cit., p. 109). Such a conclusion appears 
to me a petitio principii in regard to the relation between the 
nuclear and the cytoplasmic organizations, and that between "pre- 
formed" and "epigenetic" qualities in the cytoplasm;^ and this 

^Cf. Boveri, 1902; Wilson, 1904. 

Experimental Studies on Germinal Localization. 259 

question Is one to be answered, not by a priori considerations, but 
by observation and experiment. The facts determined by both 
these methods coincide in showing that the internal factors of 
cleavage are in a great number of cases so accurately adjusted to 
the morphogenic factors that they may be treated practically as 
identical with them. A highly differentiated initial cleavage-pat- 
tern is, therefore, ipso facto evidence of a high degree of initial 
cytoplasmic localization ; and the fact that the form of cleavage 
may be artificially altered without affecting the end result is in 
no manner opposed to this conclusion (as is pointed out in my 
nemertine paper at p. 455). 

I cannot better express the general conclusion which the facts 
seem to me to justify than by citing the following statement from 
Fischel's able general discussion ('03, p. 734) . "Der Unterschied 
zwischen den verschieden Eiarten ist demnach nur ein gradueller, 
in einzelnen Fallen vielleicht ein graduell sehr erheblicher, aber 
doch kein essentieller. Ueberall ist das Griindprincip der (nor- 
malen) Entwickeliing Mosaikarheit, und die besondere Unter- 
scheidung einer Gruppe von Eiern als 'Mosaik'-Eier ist nur in 
dem Sinne zulassig, als bei ihnen die Mosaikarheit besonders in 
Erscheinung tritt; Mosaikeier sind jedoch in gewissem Sinne auch 
alle iibrigen Eier. 

"Im Besondern ist aber noch betont, dass wohl stets niir die 
Primitivorgane des Embryo (materiell) in der Eizelle praform- 
irt enthalten sind, und dass — ganz besonders wohl bei den sogen- 
annten Regulationseiern — die materiallen Substrate fiir die Dif- 
ferenzirung der specialleren Organe wahrscheinlich iiberall erst 
wdhrend der spdteren Entwickelungsstadien gebildet werden.^ 
Im Verlaufe der Entwickelung werden stets neue und mannig- 
fache Komplikationen (in erster Linie wohl durch Stoftwechsel- 
processe, dann durch Lagebeziehungen u. a. m.) gesetzt, durch 
welche erst bestimmte Zellgruppen des Keims, und zwar vorwie- 
gend durch materielle Umwandlung oder Beimischung oder Dif- 
ferenzirung nach bestimmten Richtungen hin specificirt werden ; 
gerade dadurch aber verlieren sie die ihrem Miittergewebe friiher 
zugekomine Fdhigkeit mehr als eben nur jene specifischen Differ- 

^Cf. Wilson, '03, p. 453. 

26o Edmund B. IVilson. 

enziriingen gegenbenen Falles zu hesorgen, imd aiif diese JVeise 
gehen Beschriinkung der Potenz tind Specialisiriing fiir Organ- 
bildung einander parallel}^ * * 

"Die hier entwickelte Auffassung betont also gegeniiber jener, 
welche den verschieden Eiarten nur eine verschiedengradige 
'Regulations fahigkeit zum Ganzen' zuerkennt, die verschieden- 
stufige Abliiingigkeit zwischen Organogenese iind Eimaterialien. 
Die Entstehung von Ganzbeziehungsweise Halbbildungen aus 
Theilstiicken des Eies muss danach nicht auf jene verschieden- 
gradige 'Regulationsfahigkeit' zuriickfiihrt werden, sondern erk- 
lart sich vor Allem aus der Abhangigkeit der Differenzirung 
von der 'Qualitat ihres Ausgangsgebildes' (Boveri), d. h. von 
dem mehr oder minder vollstandigen Schichtenaufbau des betref- 
fenden Eistiickes. Der Satz von Driesch 'Jedes Element kann 
Jedes' ist demnach nur mit dem Zusatze richtig: Vorausgesetzt, 
dass dieses Element alle zur Bildung des 'Jeden' nothwendigen 
(im Ei vorgebildet enhaltenen) Plasmaqualitaten besitzt" (Ital- 
ics in the original) . 

I should only modify the above statement by recognizing the 
probability that in such extreme mosaic-eggs as those of mol- 
lusks or annelids the prelocalization may be much more detailed 
than Fischel admits, so that for example, the ectoblast may be 
represented very early in the cleavage, if not at the beginning, 
not by a single equipotent ectoblast-stuff, but by a number of such 
stuffs already specified for the production of various categories 
of ectoblast-cells (trochoblasts, apical cells, etc.). But admitting 
even to this degree the principles of prelocalization, self-differen- 
tiation, and mosaic development, it is still impossible to escape the 
parallel principle of correlative or dependent differentiation — i. e., 
the influence of the totality of the organism upon the development 
of the individual cells. For, however definitely specified a cell 
or cell-group may be, its behavior when isolated differs In some 
measure from that shown when in its normal relation to its fel- 
lows. The nature of the response to the change of conditions, 
as the facts show, is, however, conditioned and limited by the 
factors inherent in the cell or group. The further conclusions are 

'^Cf. Wilson, '93, p. 610. 

Experimental Studies on Germinal Localization. 261 

justified, I believe, that these factors differ in different types of 
eggs from the beginning, and that they become steadily more 
specialized and limited as the development progresses. With the 
advance of development, accordingly, the response becomes cor- 
respondingly more limited. 

This is shown by such a series as the following, including the 
sea-urchin, nemertine and mollusk. In all these the J/2 or Ya- 
blastomere produce an embryo that closes and gastrulates. In 
the nemertine or sea-urchin any of these embryos may undergo 
complete development, since the first two cleavages are symmetri- 
cal and quantitative, distributing to each cell all the elements of 
the original system. In the mollusk, however, the AB-half or 
the A-, B- or C-quadrant, though undergoing certain characteris- 
tic differentiations, is unable to produce a complete embryo, owing 
to the absence of necessary specific material contained only in the 
D-quadrant or the AB-half.^ Beyond the 4-cell stage all of the 
forms exhibit limitations of potency, not primarily due to decrease 
of size (as is proved by Zeleny's observations on the upper quar- 
tet of Cerebratulns) , but to qualitative Internal factors. Cells of 
the upper quartet, or the entire quartet, produce closed embryos 
which in the nemertine or mollusk are unable to gastrulate (again 
owing to the lack of specific material), but in the sea-urchin may 
do so provided the third cleavage does not exclude a certain 
amount of entoblast-stuff from the upper cells. The isolated pri- 
mary trochoblast of the i6-cell stage completes its predestined 
twQ divisions and differentiates typically except for slight changes 
in the relative position of the resulting cells; but the remarkable 
change of position, which in the complete embryo leads to the ac- 
curate fitting together of the rows of cilia, at first disconnected, 
to form continuous ciliated rings, fails to take place. Its sister- 
cell (i^) likewise divides and differentiates typically; but owing 

ij pointed out in the preceding paper that the failure of the CD-half to 
produce a perfect larva may not improbably be owing to the fact that owing 
to its great susceptibility, the larva is unable to sustain itself long enough to 
assume the normal conditon. It is theoretically possible that the same may 
be true of the AB-half; but the actual facts are that the latter shows from the 
first certain definite defects that do not exist in the former, the CD-larvce 
showing merely a lack of the proper proportions. 

262 Edmund B. Wilson. 

to the greater number of cells produced, it gives rise to an embryo 
that closes to form a morula- or blastula-like structure. The single 
trochoblast of the 64-cell stage of Patella, finally, accomplishes no 
more than a simple rounding out to a spherical form, without un- 
dergoing further modification of its predestined development. 
Each of the reactions in this series of forms must be considered 
as a response to the change of conditions that results from a de- 
struction of the relation of the part to the whole, and it seems to 
me the different cases must be considered as differing not in kind 
but in degree. In any one of these cases the inability to produce a 
perfect larva is due, as I believe, not to absolute lack of regula- 
tive capacity, but to lack of necessary material, which, as far as 
the experiments show, the cell is not able to manufacture anew; 
and the degree of regulative response may be considered, other 
things equal, as inversely proportional to the degree of segrega- 
tion that has taken place. Only, therefore, in a qualified sense, 
and in a more or less limited degree, can the prospective value of 
a cell be considered a function of its position.^ The sense in which 
this saying applies to the upper group of four in an 8-cell stage of 
Cerebratulus is far more limited than that in which it applies to 
a lateral group of four from the same stage {cf. Zeleny) . As ap- 
plied to an isolated primary trochoblast of Patella^ it becomes so 
limited as to be largely deprived of its original meaning. The 
same discrimination is necessary in considering the matter of dis- 
tribution of potencies in the cleavage-pattern. When, for ex- 
ample, Driesch asserts that "Furchungsmosaik brauch kein Mo- 
saik der Potenzen zu bedeuten" ('99, p. 729, and elsewhere), he 
is stating a fact that is incontestible as far as the 2- and 4-cell 
stages of the sea-urchin or nemertine egg are concerned, and which 
appears to apply to the medusa egg as far as the i6-cell stage; 
but when in a later paper he advances to the statement, "Fur- 
chungsmosaik ist kein Mosaik der Potenzen" ('02, i, p. 812), 
an assertion is made that is contrary to the results of experiment, 
not only on the molluscan egg from the beginning, but even on 
the nemertine or echinoderm, as soon as the 8-ceIl stage is reached. 
From the facts thus far determined the conclusion seems jus- 

iC/^. Wilson, '93, p. 610. 

Experimental Studies on Germinal Localization. 263 

tified that the power of an Isolated blastomere to produce a com- 
plete embryo depends upon three conditions : first, upon its vol- 
ume; second, upon the presence of all the essential elements (and 
apparently of the cytoplasmic elements) of the system, and third, 
upon the effectiveness of the regulative process. The production 
of a complete embryo involves the regrouping of these elements in 
a disposition essentially like that of an entire embryo, and I sec 
no escape from Driesch's contention that this is a typical act of 
regulation that cannot be explained without recourse to a factor 
that lies behind the primary topographical grouping of cytoplas- 
mic stuffs/ My observations on Amphioxus, the accuracy of 
which I see no reason to doubt, seem to show that this regrouping 
may be effected immediately upon the isolation of the cell, as 
would also seem to be the case in the inverted single blastomeres 
of the frog's egg observed by Morgan (though observations on 
the cleavage in this case are still lacking). In the greater number 
of cases thus far observed the cleavage-factors, and hence as I 
think we now may say, probably also the morphogenic factors, 
do not undergo immediate readjustment; and it is still quite an 
open question to what extent the cells formed in the ensuing par- 
tial cleavage undergo changes of prospective potence. But even 
though all the essential elements of the system be present, in a 
mass of sufficient volume, a failure of regulation may occur, per- 
haps owing to merely physical obstacles. As pointed out in my 
preceding paper, this is not improbably the explanation of the 
failure of the CD-half in Dentalium to produce a perfect larva. 
The ctenophore-egg is of exceptional interest in this direction; for 
there is good reason to conclude (since both the cleavage and the 
larva are bi-radially symmetrical) that the vertical cleavages — 
i. e.^ the first and second, and perhaps also the third — are not 
qualitative, yet, notwithstanding the closure of the embryo pro- 
duced by the >4 or M-blastomere, the larva remains defective. 
Driesch's explanation of a failure of the regulative process owing 
to a "rigidity" of organization or of protoplasmic texture seems 
in this case perfectly valid; but such explanation must be consid- 

^Cf. Lillie, '01, p. 269; Driesch, '02.1, '02.2; Wilson, '03,. p. 456. 

264 Edmund B. fVihon. 

ered inadequate for the cases of qualitative division reviewed 

As regards the relation between self-differentiation and depen- 
dent or correlative differentiation, our only guide must be the 
indirect evidence derived from the response of the cell to the 
change of conditions when its typical relation to the whole Is de- 
stroyed by isolation or displacement from Its normal position. 
For It Is perfectly obvious that If the "atypical" or secondary 
changes characteristic of an isolated blastomere do not take place 
In a complete embryo it is because of the relation of the cell to 
the whole of which It forms a part; and It Is this "relation" that 
renders the developing organism a unit, even in the most highly 
differentiated type. As to what this "relation to the whole" 
really Is we know practically nothing; but even though we employ 
a phrase of vague and uncertain content it Is of use as Indicating 
a unity or harmony of organization that is not destroyed by the 
secondary distribution of the factors of differentiation among 
localized centers. 

It is obvious that the differentiation even of such cells as the 
primary trochoblasts, which possess so high a degree of self-dif- 
ferentiation, must be definitely coordinated In some way with the 
development of the embryo as a whole, as Is shown for Instance 
by the remarkable manner in which the rows of cilia, at first dis- 
connected, are ultimately fitted together to form continuous rings 
In the prototroch; but It seems equally obvious that In such cases 
corrflative differentiation subsequent to division plays but a minor 
part In the Internal transformation of the cell. I may here point 
out, however, that the lessened Inequality of division so frequently 
observed In the Isolated blastomeres is posibly an indication of 
regulative response on the part of the internal factors of cell-dif- 
ferentiation. It is clear that the position of the spindle — and 
hence the character of the ensuing division — is definitely cor- 
related with the segregation-pattern ; and in the moUuscan egg 
many, probably all, of the earlier unequal divisions are qualitative 
In character. It is, therefore, a fair hypothesis that in these cases^ 

^The unequal division of teloblasts shows that the statement should not be 
made general without further evidence. 

Experimental Studies on Germinal Localization. 265 

the inequality is caused by, or at least correlated with, a preceding 
segregation of different materials in the cell before division. 
Hence it is an interesting fact that all the typically unequal divi- 
sions of the normal development show a tendency to become less 
unequal upon isolation of the cell. This has been observed in 
the first division of the isolated M> of the 3^-micromeres, of the 
i' cells, of the 5^-micromeres, the yV-macromeres, and in the 
CD ^-blastomere of Dentaliiim (where It is expressed by a re- 
duction in the size of the polar lobe), all of which are qualitative 
divisions. This may be explicable as a result of relatively simple 
physical conditions, but it is at the same time not improbably an 
expression of a tendency for the segregation to recede, as it were, 
towards a less definitely localized condition. The possibility is 
thus suggested that the segregative process in the cells when in 
their normal position in the whole embryo may, even in relatively 
late stages, be in some measure influenced by their relation to their 
fellows or to the whole. I believe that important light may be 
thrown on this question by an accurate comparison of the later 
development of isolated blastomeres that vary in this respect. 
The most important question remaining is whether after com- 
plete segregation and isolation of specific cytoplasmic stuffs has 
been once effected by qualitative division the missing materials 
may be restored by regulative metabolic processes. Such remark- 
able facts as those determined in regard to the regeneration of 
the lens in Triton, or Spemann's hardly less striking results on the 
formation of double-headed monsters in the same animal, leave 
no doubt that specific cell-characters may, within the limits of the 
germ-layers, be very widely altered through a response to a local 
defect, or to a change as simple as a mere mechanical alteration 
of form in the growing mass ; and the facts of regeneration even 
seem to show that one of the differentiated primary germ-layers 
may produce structures which in the typical development arise 
from a different layer. If the hypothesis of formative cytoplas- 
mic stuffs be valid there seems to be no escape from the conclu- 
sion that in such cases the necessary formative stuffs may be form- 
ed anew. But if the potentiality of the cytoplasmic system be 
primarily given in the nuclear organization, and if this be the 

266 Edmund B. Wilson. 

primary determining source of the initial cytoplasmic localization 
in the unsegmented egg, this presents no insuperable difficulty. 
It is obvious, however, that this question is one not for specula- 
tion but for further experiment. 

Zoological Laboratory, Columbia University. 
March 29th, 1904. 


BovERi, Th., '01, I. — Ueber die Polaritat des Seeigeleies: Verb. phys. med. 

Ges. Wiirzburg, N. F., XXXIV. 
Id., '01, 2. — Die Polaritat von Ovocyte, Ei und Larve des Strongylocentrotus 

lividus: Zool. Jabrb.. Anat. Abth., XIV. 
Child, C. M., '00. — The Early Development of Arenicola and Sternaspis: 

Arch. Entv^rm., IX. 4. 
CoNKLiN, E. G., '97. — The Embryology of Crepidula: Journ. Morph. XIII, 1. 
Id., '98. — Cleavage and Differentiation: Wood's Holl Biol. Lectures, 1898. 
Id., '99. — Protoplasmic Movement as a Factor in Differentiation: Ibid., 1899. 
Id., '02. — Karyokinesis and Cytokinesis, etc.: Journ. Acad. Nat. Sci., Phil., 

II Sen. XII, I. 
Crampton, H. E., '96. — Experimental Studies on Gasteropod Development: 

Arch. Entwm., III. 
Driesch, H., '92, I. — Der Werth des beiden ersten Furchungszellen in der 

Echinodermenentwicklung: Zeitschr. w^iss Zool LIII. 
Id., '92, 2. — Ueber einige allgemeine Fragen der theoretischen Morphologic: 

Zeitschr. wiss Zool. LV. 
Id., '93. — Ueber einige allgemeine entwicklungsmechanische Ergebnisse: 

Mitth. zo51. Station Neapel, XI, 2. 
Id., '94. — Analytische Theorie der organischen Entwickelung: Leipzig. 
Id., '95. — Zur Analysis der Potenzen embryonaler Organzellen: Arch. Entwm., 

Id., '96. — Betrachtungen iiber die Organization des Eies und ihre Genese: 

Arch. Entwm., IV. 
Id., '99, I. — Die Localisation morphogenetischer Vorgange: Arch. Entwm., 

Id., '99, 2. — Resultate und Probleme der Entwickelungs physiologie der Thiere: 

Merkel u. Bonnet, Ergebn., VIII. 
Id., 'go. — Die isolirten Blastomeren des Echinidenkeimes: Arch. Entwm., X, 

2, 3- 
Id., '02, I.— Neue Antworten und neue Fragen: Ergebnisse, Merkel u. Bon- 
net, XI. 
Id., '02, 2. — Neue Ergangungen zur Entwickelungsphysiologie des Echiniden- 
keimes: Arch. Entwm., XIV, 3, 4. 

Experimental Studies on Germinal Localization. 267 

Id., '03. — Drei Aphorismen zur Entwickelungsphysiologie ji'ingster Studien: 

Arch. Entwm., XVII, i. 
FiscHEL, A., '97. — Experimentelle Untersuchungen am Ctenophorenei. i: 

Arch. Entwm. VI, i. 
Id., '98.— Ibid., II-IV: Ibid, VII, 4. 
Id., '03. — Organbilding und Differenzirung: Ibid. XV. 
Heath, Harold, 99. — The Development of Ischnochiton: Z06I. Jahrb., Anat. 

Ont, XII. 
Heider, K., '00. — Das Determinationsproblem: Verb. d. Zool. Ges. 1900. 
Hertwig, O., '92. — Urmund und Spina bifida: Arch. mikr. Anat., XXXIX. 
Id, '93.— Ueber den Wert den ersten Furchungszellen flir die Organbildung 

der Embryo: Ibid, XLII. 
Id., '98. — Die Zelle und die Gewebebe, II: Jena, 1898. 
Korschelt and Heider, '02. — Lehrbuch der vergleichenden Entwicklungs- 

geschichte, Allgemeiner Theil: Fischer, Jena. 
KowALEWSKY, A., '83. — Etude sur I'embryogenie du Dentale: Ann. Mus. 

d'Hist. Nat. de Marseille. Zool. I, 7. 
Lacaze-Duthiers, '57. — Histoire de I'organisalion et du developpement du 

Dentale: Ann. Sci. Nat. IV, Ser. VI, VII. 
Lillie, F. R., '99. — Adaptation in Cleavage; Wood's HoU Biol. Lectures: 1899. 
Id., '01. — The Organization of the Egg of Unio: Journ. Morph. XVII, 2 
Maas, O., '01. — Experimentelle Untersuchungen iiber die Eifurchung: Sitzb. 

Ges. Morph. u. Phys. Miinchen, 1901. 
Id., '03. — Einfiihrung in die experimentelle Entvvickelungsgeschichte. Wies- 
Patten, W., '85. — The Embryology of Patella: Arb. Zool. Inst. Wien, VI, 2. 
Robert, '02. — Recherches sur la developpement des Troques: Arch. Zool. 

Exp., 3e Serie, Vol. X. 
Roux, W., '85. — Ueber die Bestimmung der Hauptrichtungen des Frosch- 

embryo im Ei, etc.: Gesammelte Abhandlungen, II, No. 20. 
Id., '88. — Ueber die kiinstliche Hervorbringung "halber" Embryonen, etc.* 

Ibid, II, No. 22. 
Id., '93, I. — Ueber Mosaikarbeit und neuere Entwickelungshypothesen: Ibid, 

No. 27. 
Id., '93, 2. — Ueber die Specification der Furchungszellen: Ibid, No. 28. 
Id., '95. — Gesammelte Abhandlungen, II, No. ^,2,: Nachwort. 
Id., '95. — Ueber die verschiedene Entwickelung isolirter erster Blastomeren: 

Arch. Entwm. I. 
Id., '03. — Ueber die Ursachen der Bestimmung der Hauptrichtungen des 

Embryo im Froschei: Anat. Anz., XXIII, 4-7. 
Samassa, p., '96. — Studien iiber den Einfluss des Dotters auf die Gastrulation, 

etc.; II. Amphibia: Arch. Entwm., II. 
Whitman, C. O., '93. — The Inadequacy of the Cell-Theory of Development: 

Journ. Morph., VIII. 
Wilson, E. B., '92. — The Cell-Lineage of Nereis: Journ. Morph., VI. 
Id., '93. — Amphioxus and the Mosaic Theory of Development: Journ. Morph., 


2 68 Edmund B. JrUson. 

Id., '94. — The Mosaic Theory of Development; "Wood's Holl Biol. Lect., 

II. 1893. 
Id., '96. — On Cleavage and Mosaic Work: Arch. Eiitwm., III. 
Id., '03. — Experiments on Cleavage and Localization in the Nemertine Egg; 

Arch. Entwm., XVI, 3. 
Id., '04. — The Germ-Regions in the Egg of Dentalium: Journ. Exp. Zool. I, i. 
Y.\TSU, N., '04.— Experiments on the Development of Egg Fragments in 

Cerebratulus: Biol. Bull, VI, 3- 
Zeleny, C, '04. — Experiments on the localization of developmental Factors in 

the Nemertine-Egg: Journ. Exp. Zool., I, 2. 
ZqjA, R., '95. — Sullo sviluppo dei blastomeri isolati delle uova di alcune 

meduse; Arclr. Entwm., I, II. 




(Six Figures.) 


(From the physiological laboratory of Leland Stanford, Jr., University.) 

I. The Rate of Propagation of the Nervous Impulse in the 
Ventral Nerve-Cord. 

The measurements of the rate of propagation of the nervous 
Impulse were made in the fall of 1902. Their publication has 
been delayed with the v- iew of obtaining large centipedes from the 
tropics for the work, so as to exclude a possible source of error in 
the measurements when smaller specimens are made use of. The 
attempts to obtain larger centipedes than those available here in 
California have not proved successful, and the further work must 
therefore be postponed till more favorable material becomes avail- 

The structure and relations of the central nervous system of 
the centipedes and millipedes are essentially the same as in the 
annelid worms. Each segment Is provided with a pair of ganglia, 
which are connected by transverse commissures and by longitudi- 
nal commissures with the neighboring anterior and posterior pairs 
of ganglia. This nerve-cord is situated ventral to the gut. In 
the anterior or head segment It Is connected by a commissure on 
either side of the oesophagus with the supra-cesophageal ganglion 
or "brain." 


A. J. Carlson. 

The method of measuring the rate of propagation of the nerv- 
ous impulse through this nerve-cord was essentially the same as 
that employed by Dr. Jenkins and myself in the similar work on 
the ventral nerve-cord of worms. ^ The preparation and arrange- 
ment of the animal for the experiment are shown in the diagram 
in Fig. I. The centipede was placed with its dorsal side next 
to the platform or removable floor of the moist-chamber and se- 

E / El 

I I /^ 


Fig. I. 

Diagram illustrating the method of measuring the rate of propagation of 
the nervous impulse in the ventral nerve-cord of centipedes. A, friction wheel; 
B, pins fixing one end of the reacting portion to the platform; E, electrodes; 
L, recording-lever; P, platform or floor of moist-chamber. 

■cured to the board by pins, care being taken not to injure the nerve- 
cord. In Scolopendra morsitans and Scolopocryptops sexpinosus 
three or four segments are sufl'icient as reacting or contracting 
portion, in the long and very slender centipede Himantarium taen- 
iopse"^ eight to ten segments must be used, while in the millipede 
{Jules sp.), in which the actual lengthening or shortening of 
any part of the body is very slight, eight to ten segments must be 
employed in order to give sufl'iclent amplitude to the excursion of 
the recording-lever. The segment next to the reacting portion 
was fixed to the board by means of two pins in the manner shown 

ijenkins and Carlson, Journal of Comparative Neurology, XIII, p, 259, 1903. 

-These centipedes w^ere identified for me by Mr. R. V. Chamberlin, of Cor- 
nell University. The centipede Stylolccinus, sp., made use of in studying the 
reflexes, was identified by Mr. R. E. Snodgrass. of Stanford University. 

Physiology of Ventral Nerve Cord of Myriapoda. 271 

In Fig. I, so that the contractions of the body anterior to this 
point could not be communicated to the lever. The freeing of 
the nerve-cord for the application of the distal and proximal elec- 
trodes is a very difficult undertaking, and in no instance was it 
done as completely as indicated in Fig. i, especially in the slender 
Himantariiim, in which the nerve-cord is correspondingly slender, 
and in the millipede, in which the dissection is rendered difficult 
by the very thick chitenoid epidermis. The dissection for the 
proximal electrodes was in every case made at least two or three 
segments from the reacting portion of the body, to avoid escape 
of the current directly to the reacting musculature. In Himan- 


Fig. 2. — Scolopendra. 

Tracings of the contraction of the posterior segments on stimulation of 
the cord at the distal and the proximal points. Length of cord, 5cm. Trans- 
mission time of the impulse, 0.02 sec. Rate, 2.50m. per sec. Time, 100 d v. 
per sec. 

tarium six to ten segments were allowed to intervene between the 
point of stimulation and the reacting portion. 

No anaesthetics were used, but prior to fixing the animal to 
the platform the head segment, including the cerebral ganglion, 
was usually removed. 

The posterior or tail segments of the decapitated centipede 
which has been fixed to the board and prepared in this manner 
usually become quiescent after a few minutes, and remain quies- 
cent during the intervals between the stimulation of the cord, 
provided the tension from the recording-lever is not too great. 


A. J. Carlson. 

When the tension due to the weight of the lever is considerable 
the segments are kept in constant motion until exhausted. And 
the same is true if the segmental appendages or legs are able to 
reach or touch any object. The contact of the legs with any 
solid object evidently starts reflex movements of locomotion, and 
for that reason the preparation does not become quiescent until 
nearly exhausted when fixed to the platform ventral side down 
so that the ambulatory appendages are in contact with the board. 
When the anterior end of the centipede serves as the reacting 
portion the reflex restlessness is much greater than when the pos- 
terior segments are employed. This is true whether the head seg- 
ment is removed or not. The measurements of the rapidity of 
conduction of the postero-anterior impulses in the cord by the pres- 
ent method is therefore attended with greater diflliculties than the 

Fig. 3. — Scolopaidra. 

Tracings of the contraction of the posterior segments on distal and proxi- 
mal stimulation of the cord. Length of cord, 6 cm. Transmission time of the 
impulse, 0.022 sec. Rate, 2.70 m. per sec. Time, 100 d. v. per sec. 

measurement of the antero-posterior rate. In the millipede the 
union of the segments admits of only slight elongation and con- 
traction of the body, but the body may be coiled by contraction 
of the ventral muscles in the segments. The amplitude of this 
contraction is much greater in the posterior than in the anterior 

Physiology of Fentral Nerve Cord of Myriapoda. 273 

portion of the body, and for that reason the postero-anterior rate 
of the nervous Impulse cannot very well be determined with the 
present method. 

In the centipedes Scolopendra and Scolopocryptops a single in- 
duced shock of moderate intensity applied to the nerve-cord either 
at the anterior or at the posterior end of the body produces con- 
traction of every segment in the body. In the work on these ani- 
mals the break induced shock was therefore used as the stimulus. 
This reaction to the single induced shock is not obtained in the 
long and slender centipede Himantarium or in the millipede. In 

Fig. 4.— Scolopendra. 

Tracings of the contraction of the anterior segments on proximal and dis- 
tal stimulation of the cord. Length of cord, 4.5 cm. Transmission time of the 
impulse, 0.03 sec. Rate, 1.50 m. per sec. Time, 100 d. v. per sec. 

Himantarium a single induced shock ev^en of very great intensity 
applied to the anterior or posterior end of the nerve-cord does 
not always produce a contraction that extends over the whole ani- 
mal. The contraction extends further from the point of stimu- 
lation the stronger the induced shock, but rarely from one end of 
the animal to the other. When the cord is stimulated with three 
or four weak induced shocks that follow one another in rapid 
succession the contraction involves every segment in the body. 
In the experiments on this centipede short series of the interrupted 

274 ^- J- Carlson. 

current was therefore used as stimuli. A single induced shock 
applied at one end of the nerve-cord of the millipede Jules pro- 
duces progressive movements of the ambulatory appendages or 
legs from the point of stimulation to the opposite end of the ani- 
mal, but the contraction of the muscles moving the body seg- 
ments is confined to the immediate vicinity of the point of stimu- 
lation; but a short series of the tetanizing current produces con- 
traction of these muscles in all the segments of the body. A 
similar condition was found by Dr. Jenkins and myself to obtain 
in the marine annelid Aphrodite, in which a single induced shock 
applied to the ventral nerve-cord produced contraction of the 
muscles that move the setae, but a tetanizing current was required 
to produce contraction of the muscles moving the segments. It 
is therefore probable that the nervous mechanism of the setae in 
Aphrodite and of the legs in the millipede Is less complex and 
more readily excited than is the nervous mechanism in connection 
•with the muscles that move the segments. If one of the setae in 
the worm and one of the legs of the millipede could be used for 
raising the lever and the rapidity of transmission of the impulse 
In this nervous mechanism thus measured, it would undoubtedly 
be found to be several times greater than that in the nervous 
mechanism to the segmental muscles. 

The character of the records produced by the contraction of 
the reacting portion on stimulation of the nerve-cord may be 
gathered from the typical tracings reproduced in Figs. 2 to 6. 
Only the first part of the tracings showing the latent period and 
the amplitude of contraction is given, as these are the only points 
with which we are concerned. In the records from the millipede 
(Fig. 6) the rising curves represent the gradual bending ventral- 
wards of the reacting portion, the movements of each segment 
fusing into one, apparently continuous, contraction. Each stimu- 
lation of the cord by a tetanizing current of short duration 
usually produces but one such movement. The records from the 
centipedes are more irregular from the fact that each stimulation 
of the cord usually starts a series of movements or rather con- 
tractions and relaxations which may last for a minute or two in 
the fresh preparations. 

Physiology of Ventral Nerve Cord of Myriapoda. 275 

Because of the very complex nature of the muscular part of 
the preparation the character of the curves, that is, the rapidity 
and the amplitude of the contraction is not a very accurate guide 
in determining the admissability of individual records. For ex- 
ample, two successive tracings produced by stimulation of the 
cord at the distal or at the proximal point may show great diverg- 
ence in the amplitude of the contraction and yet exhibit the same 
latent period or they may be nearly identical in the amplitude 
and rapidity of the contraction and yet show a difference in the 
latent period of from 15 to 25%. The tracings that showed a 
greater difference in the amplitude of the contractions than is 
exhibited by the records in Fig. 3 were usually excluded. 

Fig. 5. — Himantarium. 

Tracings of the contraction of the posterior segments on distal and proxi- 
mal stimulation of the cord. Length of cord, 14 cm. (120 segments). Trans- 
mission time of the impulse, 0.52 sec. Rate, 27 cm. per sec. Time, 50 d. v. 
per sec. 

Of the centipedes worked on the best preparation for these ex- 
periments is obtained from Scolopendra. The largest specimens 
yield a length of nerve-cord between the distal and the proximal 
points of stimulation of from 5 to 6 cm. This centipede is rela- 
tively stout and the reacting segments amply able to lift the re- 
cording-lever. Himantarhim is more than twice as long as Scolo- 
pendra, but is so slender that it is even difficult to fix the specimen 
to the platform without injuring the nerve-cord with the pins. 
For the experiments on this centipede the recording-lever had to 
be very light. 


A. J. Carlson. 

It was stated that the point of application of the proximal 
electrodes to the cord was always three or more segments distant 
from the reacting portion. This was done with two ends in 
view, namely, to prevent escape of the current directly on to 
the muscle and to prevent errors in the measurements from stimu- 
lation of a more direct nervous mechanism on proximal than on 
distal stimulation. In the annelids the cell bodies of the motor 
neurones to the musculature of any one segment are situated in 
the ganglia of the same segment as well as in the ganglia of the 
adjoining anterior and posterior segments. The conditions are 
in all probability the same in the nerve-cord of the centipedes and 

Fig. 6. — Juks. 

Tracings of the contraction of the posterior segments on distal and proxi- 
mal stimulation of the cord. Length of cord, 5 cm. Transmission time of the 
impulse, 0.24 sec. Rate, 21 cm. per sec. Time, 50 d. v. per sec. 

the millipedes. Now, if the cord is stimulated in the segment next 
to the reacting portion it is probable that some of the neurones 
to the reacting musculature are stimulated directly, while when 
the cord is stimulated at a point 5 to 14 cm. further away these 
neurones are probably stimulated indirectly; in other words, there 
is probably "synapses" at the junction of the longitudinal con- 
ducting paths in the cord and the motor cells to each segment. At 
such junctions the propagation of the nervous impulse is in all 
probability retarded. If therefore the latent time in the records 

Physiology of Ventral Nerve Cord of Myriapoda. 277 

on distal stimulation includes this delay while the records obtained 
on proximal stimulation do not, it is obvious that the rate of prop- 
agation of the impulse as calculated from the latent periods of 
these records would be less than the actual. For that reason it 
would be desirable to check up the measurements on these com- 
paratively short centipedes by experiments on larger representa- 
titves from the tropics, as in larger specimens this possible source 
of error can be practically excluded. 

To give an idea of the variability of the latent time in the 
records obtained by this method, three series of experiments are 
given in detail in Tables I, IV, and V. All of the experiments are 
summarized in Tables II, III, V and VII. The character of the 
tracings has already been referred to. It is amply illustrated in 

figs. 2 to 6. 


Scolopendra morsitans. Antero-posterior. Detail of experi- 
ment No. 2, Table II, October 17, 1902. Temperature, 16° C. 


Distal. Proximal. 

0.045 0.028 

0.047 ^-^^5 

0.048 0.027 

0.047 0.026 

0.045 0.025 

0.047 °-°^5 

Average. . . 0.046 0.026 

Transmission time 0-02 sec. 

Length of nerve-cord 5 ^"^• 

Rate 2.50 m. per sec. 

Summary of the measurements of the antero-posterior rate in 
the nerve-cord of Scolopendra (No. 1-8) and Scolopocryptops 
(No. 9-13). The length of nerve-cord involves from 13 to 17 

278 A. J. Carlson. 

No. of pairs Transmission Length of cord 

of records. time in sec. in cm. Rate in cm, 

I 8 0.020 5 2.50 

2 6 0.020 5 2.50 

3 II 0.030 6.5 2.16 

4 13 0-023 5 2.15 

5 4 0.025 ^-5 2.60 

6 ..... . 2 0.020 5 2.50 

7 3 0.019 6 3.15 

8 2 0.026 5 1.94 

9 4 0.015 5 3.33 

10 4 0.015 4 2.64 

II 3 0.016 4 2.40 

12 8 0.017 4.5 2.60 

13 3 0.024 6 2.46 

Mean rate 2.50 m. per sec. 


Summary of measurements of the postero-anterior rate in the 
nerve-cord of Scolopendra. 

No. of pairs Transmission Length of cord 

of records. time in sec. in cm. Rate in cm. 

I 5 0.040 6 1.50 

2 4 0.032 4.5 1.40 

3 4 0.040 7 1.75 

4 3 0.037 4 i-o8 

5 4 0.040 6 1.50 

Mean rate i .40 m. per sec. 


Himantarium taeniopse. Antero-posterior. Detail of experi- 
ment No. 2, Table V, November 5, 1902. Temperature 18" C. 

Physiology of Ventral Nerve Cord of Myriapoda. 279 

Distal. Proximal. 


0.48 0.1 1 

0.51 0.13 

0.40 0.10 

0.42 0.09 

0.46 O. II 

0.45 0.09 

0.45 0.13 

0.43 0.13 

0.45 O.I I 

0.46 0.13 

Average. . . 0.45 o.i i 

Transmission time 0-34 sec. 

Length of cord ( 100 segm.) 10 cm. 

Rate 26.4 cm. per sec. 


Summary of the measurements of the antero-posterior rate in 
the nerve-cord of Himantarium. 

No. of pairs Transmission Length of cord 

of records. 

time in sec. 

in cm. 

Rate in cm. 

I . 



^ 100 segm.) 





[ 100 segm.) 





^ no segm.) 





1 15 segm.) 





^120 segm.) 






125 segm.) 


Mean rate 



per sec. 

28o A. J. Carlson. 


Jules sp. Antero-posterior. Detail of experiment No. i, Table 

VII, November i6, 1902. Temperature 16" C. 

Distal. Proximal. 

0.40 0.16 

0.37 0.18 

0.37 0.17 

0.36 0.14 

0.36 0.15 

0.37 0.14 

0.39 0.13 

0.37 0.12 

0.40 0.16 

0.40 0.12 

0.42 0.16 

0.44 0.15 

Average. . . 0.38 0.15 


Transmission time 0-23 sec. 

Length of nerve-cord 6 cm. 

Rate 25.8 cm. per sec. 

Summary of the measurements of the antero-posterior rate in 
the nerve cord of the millipede Jules. The length of cord involves 
32 to 37 segments. 

No. of pairs Tra 
of records. t 

I 12 

2 15 

3 4 

4 2 

5 5 

6 7 

7 6 

8 6 

9 7 

ID 2 

Mean rate 20 cm. per sec. 


Length of cord 

m sec. 

m cm. 


ate m cm 





















18. 1 

Physiology of Vetitral Nerve Cord of Myriapoda. 281 

The rapidity of propagation of the antero-posterior nervous 
impulse in the cord is the same in the two centipedes Scolopendra 
and Scolopocryptops. These two centipedes are also closely alike 
in the number of segments and in the swiftness of their reactions 
and movements. The rate is lower than one might have expected, 
judging by the quick movements of these animals. While it is 
higher than the rate in the ventral nerve-cord of some of the 
worms, it is only about one-half that in the nerve-cord of the higher 
marine annelids Glycera, Eunice and Bispira (one of the Sahel- 
lidae) . 

The great difference between the rate in Scolopendra and Scolo- 
pocryptops on the one hand and that in Himantarium on the other 
is probably due to a greater number of "synapses," that is, a great- 
er complexity of the conducting path in the cord of the latter. Him- 
antarium exhibits a much greater segmental independence than do 
the other two centipedes. In Himantarium the progression of the 
contraction from the point of stimulation is slow enough to be ob- 
served by the eye, while in Scolopendra every segment of the body 
seems to contract at the same time on stimulation of the nerve-cord 
at any one point. In view of the relatively low rate even in 
Scolopendra it seems to me probable that the conducting path in 
the cord is not made up of a system of uninterrupted nerve-fibers, 
although it is evidently less complex than the corresponding con- 
ducting path in Himantarium. 

The rate in the nerve-cord of the millipede is the lowest of all, 
or only 20 cm. per sec. This is only one-third that of the lowest 
rate recorded in the nerve-cord of the annelids, namely in the 
leech (56 cm. pr sec), and in the marine worm Aphrodite (55 
cm. per sec). The reactions and movements of Jules are also 
much slower than those of Scolopendra or Scolopocryptops. 
From the fact that the rate of conduction of the impulse in the 
nerve appears to stand in direct relation to the rapidity of the 
processes of contraction in the muscle supplied by the nerve, ^ it 
seems probable that the difference in the rate in Scolopendra and 
Jules is not solely apparent and due to the greater complexity of 
the conducting path in the latter animal. 

^Carlson, American Journal of Physiology, 1904. IX, p. 401. 

282 A. J. Carlson. 

A comparation of Tables II and III leaves no doubt that in 
Scolopendra the rapidity of conduction of the impulse through 
the cord is greater in the antero-posterior than in the postero-an- 
terior direction. A similar condition exists in the case of the 
spinal cord of the California Hagfish {Bdellostoma) and there 
are indications of the same condition in the spinal cord of the 
snake/ In the annelid Glycera, on the other hand, the rate in the 
ventral nerve-cord is the same in both directions. ■ It is difficult 
to understand how this difference in the rate of conduction of the 
postero-anterior and the antero-posterior nervous impulses has 
come about in the course of development. For the preservation 
of the individual it would seem that a rapid transmission of the 
nervous impulse is just as essential over the sensory part of the 
reflex arch as over the motor part. 

II. The Reflex Functions of the Ventral Nerve-Cord and the 
Segmental Ganglia. 

The great difference in the rate of propagation of the nervous 
impulse in the cord of Scolopendra and Himantariiim lead to the 
study of the reactions and locomotions of these animals under 
natural conditions as well as of the reflexes exhibited after sever- 
ance of the head segment, together with the supra-oesophageal 
ganglion or "brain," in order to determine whether these ani- 
mals exhibit other differences in conformity with the difference 
in the rate of the nervous impulse. 

Himantarium has two modes of locomotion, namely, by means 
of its legs and by means of series of contraction waves passing 
from one end of the. body to the other exactly as in the worms. 
These movements are so identical with those of the worms that 
the muscular mechanisms are probably the same or at least simi- 
lar. The centipede works its legs at the same time that it resorts 
to the other method of getting over the ground. The worm 
method of locomotion comes into play only when the anmal is in 
a hurry to get away from an enemy. It is made use of with 

^Carlson, Archiv fiir die gesammte Physiologic, 1904, Ci, p. 231. 
-Jenkins and Carlson, loc. cit. 

Physiology of Ventral Nerve Cord of Myriapoda. 283 

equal adaptation in moving either forwards or backwards, just 
as in the worms. In Scolopendra and Scolopocryptops the legs 
are the exclusive means of locomotion, whether the progression is 
hurried or slow. The chitenoid epidermis attains also a greater 
development in these centipedes. 

Himantarium moves backwards or forwards with equal facility 
and rapidity. When at rest and touched anteriorly it runs back- 
wards; on being touched posteriorly it proceeds forwards. Scolo- 
pendra or Scolopocryptops does not move backwards for any 
length of time, and nev^er when making haste to escape from dan- 
ger, as their backward locomotion is much slower than their pro- 
gression. When Himantarium is beheaded its body keeps run- 
ning backwards continuously for ten to fifteen min. before it starts 
to move in either direction, while the decapitated Scolopendra 
keeps running forwards, no matter what obstacles are placed in 
its way, and it is very difficult to induce it to walk backwards, 
even after the excitation from the injury has partly subsided. It 
is therefore plain that Himantarium and related genera exhibit a 
less degree of antero-posterior differentiation than do the shorter 
and stouter centipedes. This is further shown by the fact that 
when the quiescent Himantarium, which is usually coiled up in a 
bunch, is gently disturbed by light or by touching it, the two ends 
of the animal will often be found to crawl or move in opposite 
directions at the same time, that is, the head end walks forwards, 
the hind end backwards, till the body is straightened out, when 
either end may take the lead. This was never observed in Scolo- 
pendra or Scolopocryptops. 

When Scolopendra or Scolopocryptops are decapitated by re- 
moving the anterior segment, inflicting as little injury as possible 
to the body, the body usually continues to move forwards inces- 
santly and rapidly for five to ten min., lifting the anterior three 
or four segments next to the wound high up from the ground. 
After the elapse of a few minutes the body becomes relatively 
quiescent, usually moving only when stimulated or touched. If 
placed on its dorsal side, the decapitated animal straightway 
turns over on its legs. When the posterior part of the body is 
touched, it either springs forwards or brings the anterior end of 

284 A.J. Carlson. 

the body around as if to bite, reactions identically the same as 
those of the intact animal. When these centipedes are cut in two 
in the middle the posterior half exhibits the reactions just de- 
scribed, with the exception that it does not turn over on its ventral 
side so readily when placed on its baclc, but it attempts to do so 
in every case. The number of segments may be further reduced 
without destroying the coordinating mechanism of locomotion. If 
the sections are made with a razor or a pair of sharp scissors, 
the whole body may be divided Into portions of three or four seg- 
ments in length, each portion still retaining coordination to the 
extent that it walks across the t^ble and keeps up locomotion for 
three to four minutes, but it exhibits no sense of equilibrium — 
that Is, attempting to turn over on Its ventral side when placed on 
Its back. The direction of the locomotion In these small portions 
of the body Is almost invariably forwards. The beheaded Scolo- 
pendra or Scolopocryptops live and react In this manner for three 
to four days. After the initial restlessness, evidently due to the 
stimulation from the lesion, it scarcely stirs If left undisturbed, al- 
though its excitability Is retained apparently unimpaired for 24 
to 48 hours. It does not react to light. When placed in a glass 
jar provided with sand or moist earth in one corner It usually 
comes to rest on these places rather than on the glass. 

The beheaded Hiinantariiim lives and reacts for seven to eight 
days, showing much more "spontaneous" activity than the decapi- 
tated Scolopendra. An 8 to 10 mm. long portion of the body 
usually exhibits the same reflexes and degree of coordination as 
the Scolopendra deprived of only Its head segment. A portion 
of that length walks forwards or backwards with apparently per- 
fect coordination of its legs, and it turns over on Its ventral side 
when placed on Its back, keeping up these reactions for 24 to 48 
hours after being isolated from the rest of the body. A portion 
of three segments walks in either direction, the usual tendency 
being to forward progression. A portion of five to six segments 
exhibits the equilibrium reflex in attempting to regain its natural 
position when placed on Its back. Longer portions turn over 

Physiology of Ventral Nerve Cord of Myriapoda. 285 

The loss of excitability and death of the decapitated Himan- 
tarium proceeds antero-posteriorly. When the animal is simply 
cut in two in the middle the anterior half with the head intact 
dies sooner than the posterior half. The same is true when this 
centipede has been bitten in the middle by a Scolopendra or a Scold- 
pocryptops, in which case the posterior half of the body usually 
lives for from 12 to 24 hours while the head end ceases to react to 
stimuli withm 2 to 6 hours. The poison of these centipedes is 
also fatal when introduced into their own bodies. When a Scolo- 
pendra is seized at its middle by a pair of forceps it usually turns 
about and bites the forceps, but occasionally It will bite the pos- 
terior part of its own body, and always with fatal results, the 
symptoms of the poisoning appearing in gradual loss of coordi- 
nation and power of locomotion, death following within 10 to 15 

The decapitated Stylolaemiis lives and reacts even longer than 
Himantarhim, or for 12 to 14 days. The only difference in the 
behavior of the decapitated and the intact Stylolaemiis seems to 
be the absence of the reaction to light in the former. The wounds 
of the decapitated Himantarhim and Stylolaemiis that lived for 
8 to 14 days healed in some cases completely. There was no in- 
dication of a regeneration of the lost part. The death was prob- 
ably due to starvation rather than to infection from the wound. 

When a number of specimens of Himantariiim and Scolopen- 
dra or Scolopocryptops are confined together where they can be 
readily observed, it will be seen that Himantariiim jerks back and 
makes haste to get away whenever any portion of Its body comes 
in contact with the other two centipedes. And it has good rea- 
sons to do so, as it Is an easy prey for these strong and ferocious 
centipedes. A similar but much less pronounced jerking back of 
the body is exhibited by all the centipedes studied when they come 
In contact with the bodies of other Individuals of even their own 
species, especially when the animals are much excited and mov- 
ing about rapidly, but In no case Is it as pronounced as in Himan- 
tariiim on coming In contact with the aforementioned species. 
The decapitated Himantariiim exhibits this very same reaction. 
Especially if the posterior end of the headless body comes In con- 

286 A.J. Carlson. 

tact with the centipedes, the body jerks back, and both modes of 
locomotion are usually employed in getting away. That the re- 
action is more pronounced when the posterior end of the body 
makes the contact is probably due to reduced excitability of the 
anterior segments next to the wound. The decapitated animal 
continues to react in this manner for several days. 

The decapitated Himantarium retreats from water just as the 
intact individual, but on coming in contact with other objects in 
its path it simply walks over or around them. When, however, 
a solid object, like a pencil or a pair of forceps, is moved towards 
the crawling centipede and the contact thus made, the decapitated 
animal usually retreats. When the body comes in contact with 
an object which is moving towards it, the impact is necessarily 
stronger than when the object is stationary and the centipede alone 
moving, hence the difference in the motor reaction is probably 
due to the quantitative difference in the sensory impulses. But 
the decapitated Himantarium jerks back and retreats from Scolo- 
pendra and Scolopocryptops even when these latter lie perfectly 
dormant, so that the reaction cannot be explained on that ground. 
One further possibility must be investigated before this reaction 
can be ascribed to a qualitative discrimination in the motor reac- 
tions to touch impressions on the part of the decapitated centi- 
pede. The touch impressions may namely be supplemented by 
those of temperature. I have made no measurements of the body 
temperature of these animals, and until such determinations are 
made this interesting point must be left undecided. 

Cross-section of the ventral nerve-cord in any part of the body 
destroys the coordination between the two ends of the body on 
either side of the lesion just as effectively as when the whole body 
is cut transversely and the two parts rejoined by a thread or a wire. 
The lesion does not destroy the coordinated locomotion of either 
half, but the direction of the locomotion of the anterior half may 
or may not be the same as that of the posterior half. When the 
direction is not the same, a "tug of war" ensues, in which the 
portion having the greatest number of segments or having the 
most favorable ground for contact for Its legs comes out victo- 
rious. Scolopendra usually turns about and bites its refractory 
hind body repeatedly. 

Physiology of Ventral J\ crze Cord of Myriapoda. 287 

When the milllped': Jules is cut transversely in the middle the 
coordination is destroyed in the posterior half. The anterior 
portion continues to move about for a short time but loss of co- 
ordination and death ensue within 10 to 20 min., and the same is 
true when the animal is decapitated. This animal is therefore 
not suited for the study of the reflexes and the relative indepen- 
dence of the coordinating mechanisms of the segmental ganglia. 

To recapitulate : Locoitiotion, movements to regain normal 
posture, as well as all contact reactions in the centipedes are ob- 
viously reflex movements not dependent on the cesophageal nerv- 
ous complex or "brain," as the decapitated centipede exhibits the 
same reactions and movements as the intact animal, save that it 
does not avoid light and cannot feed or make passages for itself 
in the ground. The decapitated centipede is not abnormally rest- 
less, so that any inhibitory functions can be ascribed to the cesoph- 
ageal nervous complex, nor is it quiescent to the extent that so- 
called "spontaneous" movements may be said to be wanting. The 
bending of the anterior part of the body preparatory to bite the 
object touching the posterior part is a reflex not dependent on the 
"brain." The maintenance of the body ventral side down is also 
a reflex through the segmental ganglia, the turning of the body 
to the ventral side when placed on its back probably d^iyending 
not so much on the touch impressions on the dorsal side as the 
absence of the normal touch impressions from the contact of the 
legs with the ground. The relatively great segmental indepen- 
dence of this equilibrium reflex and especially of the reflex and co- 
ordinating mechanisjtis of locomotion is shown by the fact that 
these are exhibited by any portion of the body measuring not less 
than three intact segments in length. 

The short and stout centipedes {Scolopendra, Scolopocryptops) 
exhibit a greater antero-posterior differentiation and a less degree 
of segmental independence than do the long and slender centi- 
pedes {Himantarium, Stylolaemus) . These latter centipedes re- 
tain the annelid mode of locomotion , and the transmission of the 
nervous impulse through their ventral nerve-cord is slower. 






(From the Rudolph Spreckels Laboratory of the University of California.) 

Comparatively few papers on the galvanotropic reactions of 
coelenterates have been published and so far as I know there are 
only two bearing directly on the questions here considered. The 
first is by PearP, who finds that when any, except the very strong- 
est, galvanic currents are passed transversely through hydra the 
animal contracts most strongly on the anode side so that the 
free end — which may be either oral or aboral — swings around 
and points towards the anode. The tentacles, however, behave 
differently. With weak currents only those tentacles which are 
parallel to the current lines contract, but of these the one towards 
the cathode has a tendency to contract most strongly. When the 
whole animal has become oriented the tentacles curve slightly so 
as to become concave on the side towards the cathode. The sec- 
ond observation is by Greeley and will be considered in detail 
later on. 

The tentacles and manubrium of Polyorchis penicillata, which 
occurs abundantly in San Francisco Bay during certain seasons 
of the year, furnish excellent material for the demonstration of 
galvanotropic reactions, responding to the current in some re- 
spects like the tentacles of hydra, but with greater distinctness. 
The method of experimentation consisted in cutting the medusae 

iPearl, 1901, Studies on the Effects of Electricity on Organisms. II. — The 
Reactions of Hydra to the Constant Current. Amer. Jour. Physiol, Vol. V, 
PP- 301-320. , 


F. W. Bancroft. 

in various ways and placing the pieces in a trough of sea water 
through which the galvanic current was conducted with non-polar- 
izable electrodes. The current strength varied from 25 to 200 'J. 
The responses were usually distinct with 25 '^, but became more de- 
cided as the current was increased. 

If a meridional strip passing from the edge on one side through 
center of the bell to the other edge be prepared and the current 
passed through it transversely, tentacles and manubrium turn and 
point towards the cathode (Fig. I) . A reversal of the current in- 



Fig. I. 

itiates a turning of these organs in the opposite direction, which 
is usually completed in a few seconds. This can be repeated many 
times and the tentacles continue to respond after hours of ac- 
tivity. The manubrium, however, tires sooner and fails to re- 
spond. If the strip is placed with its subumbrellar surface up- 
wards and extended in a straight line parallel to the current lines 
the making of the current causes the tentacles at the anode end to 


Fig. 2. 

turn through an angle of 180 degrees and point towards the 
cathode. The tentacles at the cathode end become more crowded 
together, reminding one of the tip of a moistened paint brush, 
and also point more directly towards the cathode (Fig. 2) . The 
experiment may be varied in still other ways by cutting smaller 
or larger pieces from the edge of the swimming bell, but the re- 
sponse is always the same. The tentacles wherever possible, and 
to a less extent, the manubrium, bend so as to point towards the 

Galvanotropic Reactions of Polyorchis. 291 

cathode. The response depends In no way upon the connection 
of these organs with the swimming bell, muscles or nerve-ring, 
for it is obtained equally well with isolated tentacles and pieces 
of tentacles. Isolated tentacles when placed transversely to the 
current lines curve so as to assume a more or less complete U- 

FlG. 3. 

shape, with their concave side towards the cathode (Fig. 3). 
When placed parallel to the current the tentacles do not curve 
(Fig. 3» «)• 

If the tentacles are relaxed the making of the current causes 
them to contract rapidly. Subsequently they turn their concave 
side towards the cathode, and remain contracted for a consider- 
able period. But if the current is continued long enough through 
the isolated tentacles a partial relaxation comes on which is sud- 
denly followed by another rapid contraction; so that we have, as 
this process repeats itself, a slow and irregular rhythmic contrac- 
tion caused as in the case of the quiescent frogs ventricle by the 
constant flow of the galvanic current. If the current is continued 

Fig. 4. 

still longer in some cases a local anodal relaxation occurs and the 
isolated tentacles then have the appearance of Fig. 4. As the 
figure shows, this relaxation is at the bend of the U in the curved 
tentacles and at the anodal end in those which were parallel to 
the current lines and did not curve. 

292 F. W . Bancroft. 

It is evident that these phenomena lend themselves very nicely 
to Loeb's^ explanation of galvanotropism, which he considers de- 
pends on similar changes in the tension of associated groups of 
muscles. The constant flow of the current brings about an in- 
crease of tension on the cathodal side of tentacles and manubrium, 
as a result of which this part is more strongly contracted than the 
anodal portion. When the tentacles are exhausted the anodal 
part may even be completely relaxed, 

Greeley- has stated that when "Gonionemus was exposed to the 
constant current, rhythmical contractions began always on the 
cathodal side when the medusa was immersed in normal sea 
water, but that the contractions began on the anodal side in acidu- 
lated sea water." A series of experiments was made on Polyor- 
chis to test its behavior in acid and alkaline sea water, but as long 
as the tentacles were sufficiently uninjured so that they responded 
at all to the current, they behaved as above described, «o matter 
what the reaction of the water. The influence of acid and alka- 
line media on the contraction of the muscles was also tested, but 
Greeley's results could not be confirmed. Usually a change in 
the reaction of the sea water made no difference, and even when 
it did the change in the electrical response was sometimes in one 
direction and sometimes in another, so that no significance could 
be attached to it. 

As a rule the muscles of a meridional strip of Polyorchis do not 
behave towards the galvanic current as described by Greeley for 
Gonionemus in normal sea water; for the place of maximum re- 
sponse is the anode. It is here that the contractions usually start 
and here that the most rapid rate of the rhythmic contractions is 
usually seen. But there is such an abundant opportunity for stim- 
ulation at secondary cathodes that I am not yet prepared to say 
that we have here an exception to Pfliiger's law. 

Berkeley, April 9, 1904. 

_ ^Loeb. J., 1897, Zur Theorie der physiologischen Licht und Schwerkraft- 
wirkungen. Pfliiger's Archiv. Bd. 66, p. 440: 

-Trelease, W., 1903. Report of a meeting of the Academy of Sciences of 
St. Louis. Science, N. S., Vol. XVIII, p. 753. 




1. Introduction 293 

2. Method 294 

3. Normal Development 205 

4. Experimental Results 298 

I. Unfertilized Egg 299 

II. Fertilization to Complete Separation of First Polar Body . . . 302 

III. First Polar Body to Complete Separation of Second Polar Body 302 

IV. Second Polar Body to Beginning of Lateral Elongation of Egg 306 
V. Elongated Egg to Completion of First Cleavage 308 

VI. Two-Cell Stage 309 

VII. Four-Cell Stage 313 

VIII. Eight-Cell Stage 315 

IX. Sixteen-Cell Stage 316 

X. Blastula 320 

5. Summary 2>22 

6. General Discussion 323 

The general problem of the localization of developmental fac- 
tors within the egg has received an important addition in a recent 
paper by Professor E. B. Wilson, giving strong experimental evi- 
dence of a progressive character in the localization of materials 
in the egg of Cerebratuhis lacteus.^ 

The present paper is a description of a similar series of experi- 
ments on the Mediterranean species, Cerebratuhis marginatus, 
carried on at Naples during April and May, 1903, at the sug- 

lExperiments on clevage and localization in the nemertine-egg. Archiv. f. 
Entw. der Organismen. Bd. 16 Heft 3, 1903, pp. 411-460. 

2 94 Charles Zeleny. 

gestlon of Professor Wilson/ The special aim of the experi- 
ments was two-fold. In the first place it was desired to throw 
some light upon the character of changes in localization which 
take place between the time of fertilization of the egg and the 
completion of the first cleavage. Since a fragment of the unfer- 
tilized egg segments as a whole while an isolated blastomere of 
the two-cell stage segments as a half, fragments of the egg taken 
at intermediate stages must yield interesting results. In the sec- 
ond place a comparative study was made of the characteristics ex- 
hibited by larvae developed from different portions of the egg 
isolated at the eight-cell stage. The three portions thus compared 
are the upper and the lower quartets obtained by a horizontal cut 
and the lateral four-cell groups obtained by a vertical cut. 

Clear results were obtained on these two points. For the first 
it is shown that there is a progressive localization of the cleavage 
factors between the time of fertilization and the completion of the 
first cleavage. For the second a definite differentiation along the 
polar axis of the egg is made out at the eight-cell stage. This 
differentiation occurs in such a way that while a lateral four-cell 
group remains totipotent, the upper and lower quartets are no 
longer so, one lacking the ability to form an enteron and the other 
the ability to form an apical organ. 

2. Method. 

The method of operation was a very simple one. The eggs 
were placed on a glass slide and the water was withdrawn until 
they were slightly flattened. The cut was made with a fine-bladed 
scalpel under a dissecting microscope. The resulting parts were 
then placed in individual dishes, where they were allowed to de- 
velop. In fragments of the undivided egg the segments of the 
sphere thus obtained retained their shape for several minutes, but 
gradually assumed the spherical form. Fragments of unfertil- 
ized eggs were fertilized after the spherical shape had been as- 

il wish to express my great obligation to Professor Wilson for invaluable 
advice during the progress of the experiments, and to the members of the 
staflf at the Naples Zoological Station for continued kindnesses. 

Localization of the Nemertine Egg. 295 

sumed. This method of cutting was found to be very successful 
for segmented eggs as well. Even at the eight-cell stage the upper 
and lower or the lateral groups of fours may be separated in a 
considerable percentage of cases without injury to the individual 
blastomeres notwithstanding the interlocking of the cells. 

3. Normal Development. 

The orientation of the egg in Cerehratiihis marginatus, as in 
C. lacteits, is made easy by the presence of a basal protuberance 
before and for some time after fertilization and later by the pres- 
ence of the polar bodies at the opposite pole. The protuberance 
is still evident when the first polar body is formed, but usually dis- 
appears at about the time of the formation of the second polar 
body. A considerable difference was noted in the ability to with- 
stand cutting at different periods. Before fertilization the egg 
could be cut very readily, an extra-ovate being formed in relatively 
few cases. After the first polar body had been formed, however, 
the texture of the protoplasm seemed entirely different, the eggs 
going to pieces in the great majority of cases immediately after 
the cut was made. Again after the second polar body had been 
formed the cutting property seemed much better, the quality of 
the cut resembling that of the unfertilized egg. However, it is 
very hard to draw any definite conclusion as to the comparative 
texture of the eggs at these different stages because the amount of 
flattening of the eggs on the slide, the sharpness of the scalpel 
and practice in handling the latter may have had a great deal to 
do with the cleanness of the cut. The general impression, leav- 
ing out as far as possible these disturbing factors, is that the 
protoplasm is much more liquid at the stage with one polar body 
than that at either the unfertilized stage or the two polar body 
stage. The maturation divisions seem therefore to be accom- 
panied by a profound change in the nature of the cytoplasm. 

When the egg is first removed from the body of the animal 
there is a large germinal vesicle. This is usually situated on the 
polar axis of the egg near the side farthest from the basal pro- 
tuberance (Fig. 3, p. 301). The outline of the germinal vesicle 


Charles Zeleny. 

Localization of the Nemertine Egg. 297 

Fig. I. 

Early Cleavage of Entire Egg of Cerebratultis marginatv^. 

A. two-cell stage; side view. B, four-cell stage; side view just after the 
completion of the second cleavage. C, four-cell stage; from upper pole, slightly 
later than B. D, eight-cell stage; from lower pole. E, eight-cell stage; side 
view; the relations of the larger upper quartet to the smaller basal quartet are 
shown in D and E. F, sixteen-cell stage; side view; the outline of a quadrant 
is indicated by the heavier line. G, twenty-eight-cell stage; side view, slightly 
from above; the first break in the rhythm of division is shown in the lagging 
behind in each quadrant of the cell of the intermediate group which had been 
derived from the basal cell of that quadrant. H, twenty-eight-cell stage; from 
lower pole. 

(The present figures, as well as all the following ones, unless otherwise 
described, were drawn from preparations with the aid of the camera lucida.) 

298 Charles Zeleny. 

soon fades away and within half an hour the only sign of it is a 
clear area which has collected near the pole opposite the pro- 
tuberance. In this clear area is the spindle of the first polar di- 
vision. The egg remains in this condition unless fertilized. In 
the latter case the first and the second polar bodies are formed in 
succession (Figs. 6, 7 and 9). The cell then elongates, exter- 
nally a cleavage furrow appears in the vicinity of the polar bodies 
and later another one at the opposite pole. These constrict the 
egg Into two equal parts. The second furrow, at right angles to 
the first, also passes through the polar bodies, and the two to- 
gether divide the egg into four equal parts with no cross furrow 
or very little indication of one. After the third cleavage, dexio- 
tropic as usual, the cells of the upper quartet are distinctly larger 
than those of the lower quartet. The cleavage goes on as a per- 
fect Illustration of the spiral type. The cells of the eight-cell 
stage give out smaller cells by leiotropic cleavages, taking place 
simultaneously. The further divisions of the cells of the six- 
teen-cell stage thus formed are not simultaneous. As in C. lac tens,- 
the cell of the Intermediate group in each quadrant which had 
been given off by the basal cell of that quadrant lags behind the 
others, so that there is a distinct twenty-eight-cell stage before the 
tardy cells divide to form the thirty-two-cell stage. The method 
of cleavage as here described is very constant for normal whole 
eggs, variations being extremely rare. 

The character of the normal pilidium is too well known to need 
any description here. The essential features of the early cleav- 
ages are given in Figure i, p. 296, and three stages of the larval 
development are shown In Figure 2, p. 300. 

4. The Experimental Results. 

0. Introduction. The experiments are described here in turn 
according to the period at which the operation was performed. 
Ten such periods are recognized: 

1. Unfertilized egg. 

2. Fertilization to complete separation of the first polar body. 

3. First polar body to complete separation of the second polar 

Localization of the Nemertine Egg. 299 

4. Second polar body to beginning of the lateral elongation of 
the egg. 

5. Elongated egg to the completion of the-first cleavage. 

6. Two-cell stage. 

7. Four-cell stage. 

8. Eight-cell stage. 

9. Sixteen-cell stage. 
10. Blastula. 

The limits of these periods are fairly well given in their titles 
and further definition is added later under each head. It may 
be stated that in every case where the operation was performed 
after fertilization the sperm had been added to the eggs approxi- 
mately half an hour after removal of the latter from the animal 
so that the first polar spindle was already in the metaphase in all 
cases before the entrance of the spermatozoon. This treatment 
gave a greater uniformity to the relations of maturation and fer- 
tilization than would otherwise have been possible and the exter- 
nal evidences of internal change as given by the polar bodies serve 
as good landmarks. The various groups of experiments include 
both the cases observed for the cleavage factors and those for the 
morphogenic factors. Those on the unsegmented egg were de- 
signed mainly to bring out the cleavage factors; those on the two 
and four-cell stages were intended both for cleav^age and morpho- 
genic factors; while those on the eight and sixteen-cell stages and 
blastulae were designed wholly for the morphogenic problems. 

I. Fragments of Unfertilized Eggs. The cuts in this case 
were made at periods ranging from a half hour to one and a half 
hours after removal from the mother animal, the eggs being al- 
lowed to lie in a dish of sea-water in the interval. Twenty-one 
specimens were operated on, the cuts being made in the three 
planes shown in Figure 3. Six of the cuts were horizontal, four- 
teen vertical and one oblique. 

The localization of cleavage factors (Figures 4A, B, C). In 
neither of the three groups was there an indication of a localiza- 
tion of the cleavage factors. The fragments segmented in the 
regular manner described for normal whole eggs. There is no 
cross furrow, or only a very short one, and the normal rhythm, 


Charles Zeleny. 

Fig. 2 (x 216). 

Normal Larva Dez'chpcd from Whole Eggs. 

A, larva of 233^ hours, showing the beginning of the archenteric invagina- 
tion. B, larva of 34 hours, showing enteron, mesenchyme cells and apical plate. 
C, larva of 49 hours, showing apical plate, mesenchyme cells, enteron, and one 
of the two lappets. 

size and position relations are preserved. The only abnormal case 
is the irregular flat plate of seven cells shown in Figure 4C. When 
therefore the cases are taken as a whole, the conclusion is very 
evident that the experiments give no indication of a localization 
of cleavage factors at this stage/ 

iJn connection with the early stages of cleavage the following subsidiary 
points are to be noted: 

1. The nucleated fragments formed polar bodies as in the whole egg, while 
none were formed in the non-nucleated fragments. 

2. No difference can be made out between nucleated and non-nucleated 
fragments as regards character of the cleavage, each group showing similar 
features as far as the limited data go. 

3. Probable polyspermy, as indicated by multiple division, was shown in 
two eggs. 

4. The direction of the cut has no influence upon the character of the 
cleavage. Of course, this necessarily follows from the conclusion as stated 
above, that no localization of cleavage factors is shown in the group. 

Localization of the Nemertine Egg. 


Diagram of the egg just after removal from the animal, showing germinal 
vesicle and basal protuberance. HH, VV, and OO indicate, respectively, the 
directions of the horizontal, vertical and oblique cuts. 

Fig. 4 (x 216). 

Cleavage of Fragments of the Unfertilized Egg. 

A, eight-cell stage, from nucleated fragment (r=4/5 of egg), obtained by a 
horizontal cut; viewed from lower pole. B, sixteen-cell stage, from non- 
nucleated fragment (^3/5 of egg), obtained by a horizontal cut; side view. 
C, seven-cell embryo, from non-nucleated fragment (=2/5 of egg), obtained 
by a vertical cut; viewed from convex side. 

302 Charles Zeleny. 

The localization of morphogenic factors. Only two fragments 
were allowed to develop Into larvae, and neither of these showed a 
sufficient differentiation for our purpose. One larva Is a solid 

Fjg. 5 (x 216). 

A, larva developed from a one-half vertical fragment of an unfertilized 
egg; age 49 hours. B, eight-cell stage of a fragment of a fertilized egg ( — 2/3 
of an egg), obtained by a vertical cut before the formation of the first polar 
body; except for the unusual cross furrows the cleavage resembles a normal 
whole one. 

ciliated mass of cells and the other, shown In Figure 5A, has an 
Internal cavity with a few free mesenchyme cells. 

II. Fertilization to complete separation of the first polar body. 
The limits include the period between the entrance of the sperma- 
tozoon and the complete separation of the first polar body (Fig. 
6). Seven eggs were operated on, but only two of these devel- 
oped beyond the two-cell stage. The one clear case bearing on 
the localization of cleavage factors (Fig. 5B) shows a typical 
whole cleavage at the eight-cell stage. The morphogenic factors 
receive no light from the one early blastula obtained. 

III. First polar body to complete separation of second polar 
body. The limits of this period are represented by the separa- 
tion of the first polar body on the one hand and of the second 
polar body on the other (Fig. 7), Nineteen eggs were operated 
on, seven by horizontal, ten by vertical and two by oblique cuts. 
The cases that bear on cleavage factors show In nearly every In- 
stance some departure from the normal whole cleavage as regards 
size, position or division rhythm of cells. The direction of the 
cut seems, however, not to influence the character of the defect, 

Localization of the Nemertine Egg. 



Fig. 6 (x 180). 

Diagram of Egg Soon After Fertilization and Before the Fai-mation of the First Polar 

The dotted line at one pole incloses a clear area containing the first polar 
spindle. The basal protuberance is still very prominent; HH, VV and OO 
represent, respectively, the directions of the horizontal, vertical and oblique 

Fig. 7 (x 180). 
Diagram of Egg with One Polar Body. 
The basal protuberance is still evident, but not as prominent as in the for- 
mer stages; HH, VV and OO represent, respectively, the directions of the 
horizontal, vertical and oblique cuts. 

at least as far as the present experiments go. The most common 
irregularity is a simple departure from the normal division rhythm 
(Fig, 8D), which in other cases is accompanied by a displacement 
of some of the cells (Figs. 8 A, B, C). An interesting plate form 
was obtained from a vertical fragment (Figs. BE, F, G). Its 
special interest comes from the fact that notwithstanding the ir- 
regularity it formed a SAvimming larva on the second day, 24 
hours after fertilization. 


Charles Zeleny. 

Localization of the Nemertine Egg. 305 

Fig. 8. 

Fragments Obtained from Eggs with One Polar Body. 

A, 8-i6-cell stage of a fragment (=14/5 of egg), obtained by a vertical cut; 
side view; variations from the normal whole in rhythm of division, in size re- 
lations and in position of blastomeres are to be noted. B, side view of same 
egg from other side after a horizontal rotation through 180 degrees. C, same 
egg viewed from lower pole. D, 8-i6-cell stage of the lower fragment ( — 6/7 
of egg), obtained by a horizontal cut; side view; variations are to be noted in 
rhythm of division and size relations of cells. Judging by the character of the 
cut, the fragment probably contained the sperm nucleus without the egg 
nucleus. E, 4-6-cell stage of fragment (=^3/5 of egg), obtained by a vertical 
cut. F, 8-cell curved plate form of same. G, lo-cell curved plate form of 
same. H, larva from fragment (=zupper 2/3 of egg), obtained by a horizontal 


Charles Zeleny. 

None of the three larvae which were allowed to develop showed 
a sufficient differentiation of parts to be of service in determining 
the localization of the morphogenic factors at this stage. The 
one figured (8H) is a large blastula, the cavity of which is filled 
with free spherical cells. The other two larvae, one of which was 
just mentioned above, both developed after extremely irregular 
cleavages, and are interesting because they show the presence of 
an extremely high power of regulation. However, because of the 
lack of data as regards their structure no definite conclusion can 
be drawn even here. 

IV. Second polar body to beginning of lateral elongation of 
the egg. The period is limited on the one hand by the complete 
separation of the second polar body from the egg and on the 
other by the division of the cleavage nucleus and the accompany- 
ing elongation of the cell preparatory to the first cell division 
(Fig. 9). Eleven eggs were operated on, four by horizontal. 


Fig. 9 (x 180). 
Diagram of Egg zvith Tzvo Polar Bodies. 

The basal protuberance has disappeared; HH, VV and OO represent, 
respectively, the directions of the horizontal, vertical and oblique cuts. 

four by vertical and three by oblique cuts. The fragments do not 
segment as wholes, but shotv very evident departures from that 
mode. However, there is a wide difference in the extent of this 
departure in different cases. As determined by the experiments, 
the range of the variation is from a possible whole cleavage, 
through cases with a slight disturbance in size and position of 

Localization of the N emertine Egg. 


cells or rhythm of division (Figs. loA, B, C) up to a case with 
an open cup-shaped blastula of a purely partial type (Fig. loE). 
The results give no definite relation between the position of the 
removed portion of the egg and the character of the resulting 
defect in cleavage. A possible instance of such a relation is shown 
in a sixteen-cell stage developed from a vertical-oblique fragment, 
which show^s a corresponding flattening of one side of the embryo 
(Fig. loC). ■ 

Fig. 10 (x 216). 

Fragments Obtained from Eggs 7vifli Two Polar Bodies Before the Lateral Elonga- 
tion of the Cell. 

A, eight-cell ( — ) stage of fragment (3=2/3 of egg), obtained by a vertical 
(slightly oblique) cut; oblique view. Note that the cells of the upper quartet 
are smaller than those of the lower (the reverse of the normal condition) and 
that one of the quadrants is behind the others in its division. B, 15-16-ceIl stage 
of fragment (=1/2 of egg), obtained by an oblique cut; side view. Note that 
the cells differ from normal whole ones in rhythm of division and size rela- 
tions. C, i6-cell stage of a fragment (=22/;^ of egg), obtained by vertical- 
oblique cut; side view. Note oblique flattening of egg. D, larva (50 hours 
old) from fragment (:r=2/3 of egg), obtained by a vertical cut; side view. 
Note solid enteron and absence of apical plate. E, open, partial blastula from 
a fragment (3=2/3 of egg), obtained by a vertical cut; oblique view from 
open side. 


Charles Zeleny. 

The localization of morphogenic factors is not particularly 
elucidated by the two larvae obtained. The one represented in 
Figure loD has a solid interior cell mass, evidently an archen- 
teric ingrowth. There is no other structure sufficiently differen- 
tiated for our purpose. 

V. Elongated Egg to completion of first cleavage. The period 
is limited on the one hand by the beginning of the lateral elonga- 
tion of the egg and on the other by the completion of the first 
cleavage. Five eggs were operated on. In every case the frag- 
ments obtained show a partial cleavage from the start, even 
though in several instances the cleavage furrow was slight at the 
time of the operation, and there was still a broad connecting band 
between the two parts of the egg. This band was in every case 
equal to one-half or more of the diameter of a blastomere of the 
two-cell stage. The two fragments from one of the eggs are 
shown in Figures 1 1 A and B at the four-cell stage. It is evident 

Fig. II (x 216). 

Fragments from Eggs Bctzveni the Beginning of Lateral Elongation and the Coin- 
pktion of the First Cleavage. 

A, four-cell stage of fragment {=111/2 of egg), obtained by a vertical cut. 
Polar bodies are attached. The second furrow is equatorial. B, four-cell 
stage from the other half of the same egg. 

that the cleavage is a partial one resembling closely that of iso- 
lated blastomeres of the two-cell stage to be described later (p. 
309) . One of the fragments has the two polar bodies still attach- 
ed, and it is evident that the second cleavage furrow is equatorial 

Localization of the N emertine Egg. 309 

and not vertical as in the whole egg. The fragments from three 
of the eggs developed into cup-shaped half blastulae again, re- 
sembling the similar embryos arising from isolated blastomeres 
of the two-cell stage. There is, therefore, at this period a defi- 
nite localization of cleavage factors. 

As regards the localization of the morphogenic factors no gen- 
eral statement can be made. The two larvae obtained did not 
show a sufficient differentiation to be of value. 

In the experiments on unsegmented eggs a study of the locali- 
zation of the cleavage factors has been the main object in view, 
the few and unsatisfactory isolated observations on larvae develop- 
ing from the fragments being incidental and subsidiary to the 
main point. In the following experiments, however, the study 
of the localization of the morphogenic factors is definitely taken 
up, the most extended series and the one yielding the most inter- 
esting results being on the eight-cell stage. The experiments on 
the localization of the cleavage factors are continued for the two- 
cell and four-cell stages. 

VI. Two-cell stage. 

I. Experiments on the localization of the cleax-age factors in 
isolated blastomeres. The blastomeres were isolated in twenty- 
eight eggs of the two-cell stage. In the majority both blastomeres 
segmented, a minority showing no cleavage of one of the parts. 
In nearly every case the cleavage could be recognized as a partial 
one corresponding with that of a lateral half of the whole egg. 
At the four-cell (8/2) stage there is a wide cross furrow and the 
cells are not in the same plane. In fact they appear very much 
as if they had been removed from the whole eight-celled embryo 
by a vertical cut (Figure 12 A, B). The different forms of cleav- 
age described for isolated blastomeres of the two-cell stage of 
C. lact-ens by Professor Wilson were found here also. Their 
characteristics are especially prominent during the eight-cell 
(16/2) and the sixteen-cell (32/2) stages (Figs. i2AtoI, 13A). 
The most numerous are the cup-shaped embryos resembling a 
geometrical half of a w^hole blastula of the corresponding age 
(Figs. 12I, 12H). On the one hand the cups are replaced by 


Charles Zeleny. 

Fig. 12 (x2i6). 

Cleavage Stages of Isolated Blastomeres of the Two-Cell Stage. 

A, four-cell stage from an isolated blastomere. Note wide cross furrow; 
also, that the blastomeres are not in one plane. B, four-cell stage. C, eight- 
cell stage. D, eight-cell stage of plate type. E, eight-cell stage of slightly- 
curved plate type; viewed from convex side. F, eight-cell stage of curved 
plate type; side view. G, eight-cell stage of very shallow cup type; viewed 
from convex side. H, eight-cell stage of shallow cup type. I, 16-32-cell stage 
of cup shaped type (i^geometrical half of normal blastula) ; side view. 

Localization of the Nemertine Egg. 3 1 1 

flat plate-like forms, there being all gradations between the em- 
bryos curved into a deep cup through those showing only a slight 
curvature up to perfectly flat plates (Figs. 12D to G and 13 A). 
On the other hand there is a similar graded series from the cup 
forms up to perfectly closed spherical half-embryos usually con- 
taining a very small cavity or none at all (Fig. 12C). It seems 
probable that these differences of form are the result of slight 
changes in the surface tension relations between the cells, as Pro- 
fessor Wilson has suggested, and this view is strengthened by 
my observation of the development of a plate form and a cup 
form from the two blastomeres of a single egg. 

2. Experiments on the localization of the morphogenic fac- 
tors. The localization of the morphogenic factors in the two- 
cell stage was not made out as fully as could have been wished. 
The larvae were killed in most cases at too early a stage to deter- 
mine the necessary difi^erentiation of organs. The two blasto- 
meres were separated in each of sixteen eggs and in thirteen larvae 
were obtained. Most of these were about 33 hours old when 
killed, only three being older than this. The thirteen individuals 
are divided into groups of similar cases in the following descrip- 

In three cases observations were made on the activity of the 
embryos, but the embryos themselves were lost during transfer- 
ence to the preserving liquid. An interesting fact in connection 
with these, and this holds also for other one-half as well as one- 
fourth embryos, is the abnormally great rapidity of rotation in 
most of the cases. 

Another group is formed by isolated blastomeres from five eggs. 
The larvae were distinguished by rapid rotation in life and by a 
dense ingrowth of cells from one pole, which entirely filled the 
blastocoele and came close up against the ectoblastic wall around 
the whole surface of the egg (Fig. 13B, C, D). No apical plate 
was made out in any of them, but in one case there was a single 
lappet (Fig. 13B). 


Charles Zelen 


Fig. 13 (x 216). 

Cleavage and Larval Stages from Isolated Blastonieres of the Tivo-Ccll Stage. 

A, nine-cell stage (rrrcup sbaped type); view from concave side. The cut 
was made at one side of cleavage plane so that the fragment included one 
blastomere plus part of the other. B, larva (age=z:33^ hours). Note single 
lappet, solid enteron and absence of apical organ. C, larva (age=:33 hours). 
Note solid enteron and absence of apical organ. D, larva (age^33^ hours). 
E, larva (age=:r33 hours). F, larva (agez=33^ hours). G, larva (age=:33 
hjours). The blastomeres were not completely separated and may have fused. 
H, larva (age=47 hours). Note apical organ and small enteron. 

In three cases the embryos resemble the five just mentioned, ex- 
cept that the archenteric mass is not as large and a slight blasto- 
coele, crescentic in vertical section, is present (Figs. 13E, F) . 
This blastocoele contains rounded and irregular mesenchyme cells. 
There is no apical organ. The larvae do not differ widely from 
the normal larva of about 24 hours though their age is 33 hours. 

Localization of the Nemertine Egg. 313 

In one of the two remaining cases the blastomeres were not 
completely separated. The result was two connected partial em- 
bryos in the early stages, which evidently later fused to form a 
single individual. The resulting larva (age 33 hours) shows a 
large blastocoele, two apical organs and a solid enteric mass grow- 
ing in at the base. The blastocoele contains free rounded cells, 
and there are no lappets (Fig. 13G). 

Finally there is the one case which was allowed to develop for 
a sufficient length of time (47 hours) to give the organs a chance 
to differentiate. The resulting larva (Fig. 13H) rotated rapidly 
in life. It has a large blastocoele, a small enteron, an apical plate 
and a thickening in the wall at the side of the mouth opening, 
probably representing the basis of the ectodermal invagination at 
this point. There are no lappets. With the exception of the small 
size of the enteron and the absence of the lappets, the larva does 
not differ widely from a normal larva. 

Summary of the results on the localization of morphogenic fac- 
tors. The larvae developed from isolated blastomeres of the two- 
cell stage do not show any constant defects except possibly as re- 
gards the lappets, organs which in C. marginatus are developed 
at a comparatively late period. Of the instances here cited only 
two can be considered as old enough to have formed the lappets. 
At any rate we must consider the larva developed from an iso- 
lated blastomere of the two-cell stage to be retarded in develop- 
ment as compared with a normal one of the same age, though this 
view does not serve to explain completely the characteristics of 
several of the larvae. 

VII. Four-cell stage. The experiments at this period come 
under two heads. In one series the segmenting egg was divided 
into two groups of two cells each, and in the other the four blasto- 
meres were isolated. 

The isolated blastomeres segment in every respect as quadrants 
of the whole egg. It will be remembered that the whole egg of 
Cerehratulus goes through a definite twenty-eight-cell period be- 
cause one of the cells of each quadrant lags behind the others in 
its division as the egg passes from sixteen to thirty-two cells (see 
Fig. iG). Correspondingly the isolated blastomere of the four- 


Charles Zeleny. 

cell stage passes through a definite seven-cell (28/4) stage. Such 
a stage is represented in Figures 14B and C. 

Fig. 14 (x 216). 

Cleavage Stages and Larva from Two-Cell Groups and Isolated Blastomcres of the 

Four-Cell Stage. 

A, seven-cell stage from fragment (two cells+) of tgg. B, seven-cell 
stage from isolated blastomere of an egg. C, seven-cell stage from other 
isolated blastomere of egg shown in B. D, larva (ager=33 hours) from two- 
cell group. Note apical plate, solid enteric invagination and large balstoccele 
with numerous free cells. E, larva (age=r33 hours) from isolated blastomere. 
Note the very large solid enteron nearly filling the blastocoele. 

Ten eggs were used for the experiments on the localization of 
the morphogenic factors. The only larva developed from a two- 
cell fragment was asymmetrical and swam in a small circle. 
Thirty-three hours after fertilization it had an apical plate and 
cilia, the beginning of the ingrowth of the archenteric mass and a 
large blastocoele containing rounded and irregular mesenchyme 
cells (Fig. 14D). There is thus no definite specification of the 
morphogenic factors in a two-cell group of the four-cell stage. 

In six cases the four blastomeres were isolated, and in five of 
them very rapidly rotating larvae resulted. The observations on 
these were made in most cases thirty-three hours after the fertili- 

Localization of the Nemertine Egg. 315 

zation. The larvae resemble very much the solid larvae of the 
isolated blastomeres of the two-cell stage, but are of smaller size 
(Fig. 14E). All the one-fourth larvs are like the one figured. 
There is a solid archenteric growth, but no sign of an apical organ. 

The larvae from isolated blastomeres of the four-cell stage, 
therefore, give no indication of a definite localization of the mor- 
phogenic factors, though they do not develop in an entirely nor- 
mal manner. The latter statement holds also for the larvae de- 
veloped from the isolated blastomeres of the two-cell stage, as has 
already been stated (p. 311). 

VIII. Eight-cell stage. The results yielded by the experiments 
on this stage are perhaps the most important of all those given. 
The group of experiments included sixty-eight four-cell groups. 
These groups were separated by a careful cut with the fine scalpel 
blade used in all the experiments. In most cases the knife blade 
passed between the cells, and the latter were entirely uninjured by 
the operation. In a few, however, the protoplasm was cut, and 
these will be mentioned in the descriptions. The operations in- 
clude a series of horizontal cuts separating the upper from the 
lower quartet, and a series of vertical cuts separating the two 
lateral four-cell groups, each of the latter containing two cells of 
the upper and two of the lower quartet. There are thus three 
kinds of four-cell groups, the larvae from which are to be com- 
pared: (i) Upper quartets, (2) lower quartets, and (3) lateral 
four-cell groups. The experiments yield a very definite and posi- 
tive result. The larva developing from the upper quartet have 
an apical organ, but no archenteron, those from the lower quar- 
tet have an archenteron, but no apical organ, while those from 
lateral four-cell groups have both apical organ and archenteron. 

The natural conclusion '4o be drawn from these results is that 
certain organ-forming materials are definitely separated by the 
third cleavage plane, and the larvae developing from the lower 
or the upper quartet have not the power of making up the lack- 
ing material. The lateral four-cell groups, howev^er, possess both 
kinds of materials and are, therefore, able to develop both archen- 
teron and apical organ, though the larvae are usually asymmet- 
rical. , 

3i6 Charles Zeleny. 

Some of the larvae are shown in Figures 15, 16 and 17, and it 
will not be necessary to describe the individual cases in detail, as 
the results are very definite and clear. The figures give charac- 
teristic types of larvae developing from the upper quartet (Fig. 
15), from the lower quartet (Fig. 16), and from lateral four-cell 
groups (Fig. 17A, B, C). Figure 17D shows a case in which 
six of the eight cells were represented, two of the lower quartet 
having been destroyed. 

IX. Sixteen-cell stage. Five eggs of the sixteen-cell stage suc- 
cessfully withstood an operation, and larvae from three of these 
were studied. 

In one egg equal upper and lower portions were obtained by a 
horizontal cut, but there was not a separate identification of them, 
and they were placed in one dish. At forty-eight hours after fer- 
tilization both resultant embryos were ciliated. They showed a 
difference in that one had ragged edges and swam in a circle, 
while the other had even edges and remained stationary. The 
embryos were lost. 

In two cases the upper four cells were successfully separated 
from the lower twelve. The two cases are taken up in turn. In 
the first one the division was very clear without injury to any of 
the cells. At forty-six and a half hours after fertilization the 
upper four cells had 'formed a small distinctly outlined spherical 
ciliated embryo, with no rotation or forward motion of the body. 
There is a distinct blastocoele containing rounded cells, a large 
apical organ and no enteron or lappets (Fig. 18C). At the same 
time the lower twelve cells have formed a ciliated rotating embryo, 
with a large solid archenteron entirely filling up the cavity of the 
blastocoele. Neither apical organ nor lappets are present. The 
two embryos thus show a very pronounced difference, the one 
formed from the upper four cells containing an apical organ and 
no archenteron, and the other, from the lower twelve cells, con- 
taining an archenteron and no apical organ. 

In the remaining case the upper four cells were separated from 
the lower twelve as before. One cell in the former was slightly 
injured, but all the cells of the latter were left in good condition. 
From the upper four cells at forty-six and a half hours after fer- 

Localization of the Nemertine Egg. 


Fig. 15 (x 216). 

Larva from U[^pcr Quartets of the Eight-Cell Stage. 

A, larva (age^44 hours) from complete upper quartet. B, larva (age 
=:4S hours) from upper quartet, with one cell injured. C, same larva rotated 
horizontally. D, larva (age^45 hours) from complete upper quartet. The 
egg already showed the cell constrictions for the next (i6-cell) division. E, 
same larva rotated horizontally. F, larva (age:=45 hours) from complete 
upper quartet. The inner cell mass does not connect with the side of the 
larva (i. e., it is free). G, larva (ager=23 hours) from complete upper quartet. 
H, larva (age:=33 hours) from isolated blastomere of the upper quartet. I. 
double larva (age=46 hours) from upper (?) quartet. Note presence of apical 
organ and absence oi enteron in Numbers A to F. 


Charles Zeleny. 

Fig. i6 (x 216). 

Larvcc from Lower Quartets of the Eight-Cell Stage. 

A, larva (age:i=46 hours) from lower (?) quartet. Note large enteron and 
absence of apical organ. B, larva (ager=46 hours) from lower quartet. C, 
same larva rotated horizontally: Note solid enteron and absence of apical 
organ. Ahte large archenteric ingrowth and absefice of apical organ in all cases. 

Fig. 17 (x 216). 
Lan'cc from Portions of the Eight-Cell Stage. 

A, larva (age=:48 hours) from lateral four-cell group. Note presence of 
both enteron and apical organ. B, same larva rotated horizontally. C, larva 
(age=r:33 hours) from lateral four-cell group. Note presence of both enteron 
and apical organ. D, larva (age^48 hours) from upper quartet plus two cells 
of lower quartet. Note three apical organs, large blastocoele, small enteron 
and two ectodermal invaginations at sides of enteron. 

Localization of the N emertine Egg. 


Fig. 18 (x 216). 

Larvce from Portions of the Egg oi tin Sixtecn-Ccll Stage. 

A, larva (agc:r=46j^ hours) from the upper four cells (one injured). 
Note presence of apical organ and absence of enteron. B, larva (age^46y2 
hours) from the lower twelve cells of the same egg. Note that egg is a solid 
mass with division into ectodermal and endodermal cells. C, larva (age 46^4 
hours) from upper four cells. Note presence of apical organ and absence of 
enteron. D, larva from lower twelve cells of the same egg; same age. Note 
solid enteron and absence of apical organ. 

tilization an elongated very actively swimming larva with a long 
apical cilium had developed. The embryo showed after stain- 
ing and mounting a well developed apical plate. The anterior 
end of the body is occupied by a blastocoele containing a few scat- 
tered free cells. The posterior end is a dense mass of cells, with 
no signs of an ingrowth of these to form an archenteron (Fig. 
18A). From the lower tweh^e cells at the same time there was 
developed a ciliated elongated embryo, with only a slight rotation, 
and no forward movement of the body. The embryo is a solid 
mass of cells, the only differentiation visible being a difference be- 
tween the cells at the two ends. Those near one end have the 
typical histological endoderm characters of the normal larva, 
while those near the other end have ectoderm characters (Fig. 

The characters of the two larvae in this case again show the 
presence of the apical-basal differentiation described for the last 

320 Charles Zeleny. 

specimen. The experiment seems to indicate that the basis of the 
apical organ is found in the four upper cells of the sixteen-cell 
stage. In connection with this result Yatsu's observations on the 
unsegmented egg of C. lacteus are interesting. He localizes the 
basis of the apical organ in a broad band just above the equator 
of the egg. 

X. Blastula stage. Successful operations were made on three 

The first one was divided by a horizontal cut into an upper part 
( = 2/3 of blastula) and a lower part ( = 1/3 of blastula). The 
orientation was made certain by the presence of the polar bodies. 
The upper part broke up into two portions, each of which at 
twenty-four hours had developed into an embryo with an apical 
cilium. At the same time the embryo from the lower one-third 
of the blastula was ciliated but had no apical organ. At forty- 
seven and a half hours the embryo from the lower one-third and 
one of the upper ones were dead. The other upper embryo has 
two apical plates, one a well developed and the other a small one, 
an invaginated ectodermal sac, a large and well developed en- 
teron, a blastocoele with free cells in its cavity, and no lappets. 
In fact, except for the absence of the lappets and the presence of 
two apical organs, it has all the characters of a typical whole 
larva (Fig. 19C; only one of the apical plates is shown). How- 
ever, at twenty-four hours, as stated above, there is a distinct dif- 
ference between the upper embryos and the lower one because of 
the presence of the apical organ in the former and its absence in 
the latter. 

A second blastula was cut into equal upper and lower parts by 
a horizontal cut, but the two were not kept separate. One of the 
halves died. The other developed all the organs of the normal 
pilidium, except the lappets. There is a large blastocoele, two 
apical organs, one in the normal position and one asymmetri- 
cally placed and not shown in the figure, and a large long enteron 
straighter than in the normal larva (Fig. 19B). 

A third blastul-a was cut into two unequal parts equal respec- 
tively to two-thirds and one-third of the blastula, by a cut of un- 
known direction. One portion, the larger one judging by the 

Localization of the Nemertinc Egg. 


Fig. 19 (x2i6). 

LarvcE Developed from Blastula Fragments. 

A, larva (age 351^^ hours) from a fragment (1^:2/3 of blastula) ; direction 
of cut is not known. B, larva (age 3514 hours) from upper or lower half of 
blastula. Note presence of both enteron and apical organ. C, larva (age3r48 
hours) from the upper 2/3 of blastula. Note presence of both enteron and 
apical organ. 

size, was the only one alive thirty-five and a half hours after fer- 
tilization. Its cilia were waving, but there was no motion of the 
animal as a whole. The body is spherical, with a large solid en- 
teron nearly filling the cavity of the blastocoele. The chief defect 
is in the absence of the apical organ, as it is too early (35/^ 
hours) for the lappets to appear (Fig. 19A). The direction of 
the cut is not known and the defect, therefore, cannot be corre- 
lated with any definite portion of the blastula. 

The experiments on blastulae give only one organ which can be 
considered as definitely specialized. The apical plate is developed 
in each of two embryos from the upper two-thirds of a blastula, 
while it is absent in those developed from the lower one-third. 
No explanation can be given of the apparently greater regulative 

322 Charles Zeleny. 


power of blastula fragments as compared with those of the eight- 
cell and sixteen-cell stages. A similar fact was noted by Professor 
Wilson for C. lacteus, and he supposes that a possibility of error 
in orientation of the blastulae may account for the result. For this 
reason I took special care in determining the orientation, and in 
two of the three cases I think there is little doubt of the correct- 
ness of the determination. 

5. Summary of results. 

1. Both nucleated and non-nucleated fragments of the unfer- 
tilized eggs of Cerebratulus marginatus segment as wholes. 

2. Isolated blastomeres of the two-cell stage segment as if the 
other blastomere were still in its place, i. e., they segment as ver- 
tical halves. 

3. Fragments obtained during the stages between the fertili- 
zation of the egg and the completion of the first cleavage show a 
progressive specification of the cleavage factors as evidenced by 
abnormalities in rhythm of division, size relations of cells and 
position of cells. After the separation of the cleavage nuclei and 
when the cytoplasm of the two cells is still widely connected, the 
two halves when cut apart may already show all the characters of 
half cleavages. 

4. Isolated blastomeres of the four-cell stage segment as 
fourths of the whole cleavage pattern. 

5. Larvae developed from the upper quartet of the eight-cell 
stage always possess an apical organ and lack an enteron, those 
developed from the lower quartet always possess an enteron and 
lack an apical organ, while those developed from lateral four- 
cell groups containing two cells of the upper and two cells of the 
lower quartet always possess both apical organ and enteron. 

6. Larvae developed from the upper four cells of the sixteen- 
cell stage lack an enteron, but possess an apical organ and blasto- 
coele. Those developed from the lower twelve cells have a large 
enteron, but no apical organ or blastocoele. 

7. Two embryos developed by a secondary division from the 
upper two-thirds of a blastula both developed apical organs. The 

Localization of the Nemertine Egg, 323 

embryo developed from the lower one-third of the same blastula 
developed no apical organ. 

6. General discussion. 

The points brought out by the present experiments are of con- 
siderable general interest. In the first place in agreement with the 
results of Professor Wilson on C. lacteus, it is found that while an 
egg fragment of an unfertilized egg segments as a whole an iso- 
lated blastomere of the two-celled stage segments as a half. In the 
intermediate stages there is a gradually increasing departure from 
a whole cleavage in the fragments as we pass from the first men- 
tioned stage to the latter. This is contrary to the statement made 
by Yatsu in his recently published paper on C. lacteus. In prin- 
ciple, however, it agrees with the progressive increase of defects 
found by him in larvae developing from fragments taken at simi- 
lar stages. 

Though the observations naturally suggest the view that there 
is a progressive localization of materials in the egg from one 
period to the other, such a conclusion does not necessarily follow 
from the experiments themselves without further data. Because, 
considering the power of regulation of the embryo shown at all 
stages studied, it must be admitted that there remains the possi- 
bility of regulation of the unfertilized fragment to form a com- 
plete whole cleavage and later a complete larva. For the earlier 
the operation be performed the greater the time which must elapse 
before the fragment divides, and consequently the greater the 
chance for regulation to a whole cleavage pattern. The experi- 
ments of Schultze on inversion of whole frog's eggs at the two- 
celled stage and the corresponding ones of Morgan on the isolated 
blastomeres of the same stage, show that the rearrangement of 
materials due to difference in specific gravity gives opportunities 
for regulation to a whole development. Observations on the nor- 
mal eggs of a great many animals during the maturation period 
show a very extensive series of streaming movements in the proto- 
plasm at this time. May not these furnish a similar opportunity 
for regulation to a whole cleavage and whole development? For 

324 Charles Zeleny. 

as the materials In the Isolated blastomere of the frog's egg are 
undoubtedly specialized so as to form a half cleavage pattern and 
a half embryo under the ordinary conditions, the re-adjustment of 
materials due to Inversion gives the necessary conditions for regu- 
lation. May not the unfertilized egg of Cerebratulus likewise 
show a localization of formative factors so that a fragment is a 
true portion of a mosaic, but needs only the conditions accompany- 
ing the streaming during the maturation stages to accomplish a 
readjustment to a whole? Evidently there Is no means of de- 
termining this point, because If we assume the possibility of a re- 
adjustment during the maturation stages, we remove our only hope 
of a direct method of deciding the question as to the presence or 
absence of developmental specification before fertilization. The 
only remaining method lies In Indirect Inferences from the ob- 
served localization of visible materials during these stages. There 
Is abundant proof of such a progressive localization during this 
time, and the conclusion that there is an arrangement of forma- 
tive materials Into a definite pattern at this period Is a natural one. 
For there can be no valid objection to the association of the two 
parallel processes of localization of visible materials and of re- 
sultant cleavage and morphogenic factors. 

But why Is there a progressive localization of the morpho- 
genic factors In the unsegmented egg, as Indicated by Yatsu's 
work, while at the same time the Isolated blastomere of the two- 
celled stage develops Into a perfect larva ? The progressive local- 
ization of cleavage factors as shown in my experiments Is naturally 
to be expected, since there is a gradual passage from a whole 
cleavage on the one hand to a half cleavage on the other. We 
may assume a gradual localization of materials controlling these 
factors, or a greater opportunity for regulation In the earlier as 
compared with the later stages,' or both, to account for the data. 
With the progressive localization of morphogenic factors, as de- 
scribed by Yatsu, there Is no such sequence. Starting with com- 
plete larvae developing from the fragments of unfertilized eggs, 
there is a gradual increase in the defects in the larvae up to the 
completion of the first cleavage. Then very suddenly, as soon as 
the cleavage is completed, there Is a return to whole larvae. I 

Localization of the Nemertine Egg. 325 

think the apparent contradictions may be explained in the fol- 
lowing way, though the purely speculative character of all the 
discussions is always to be kept in mind. 

Numerous recent observations, especially those of Lillie and 
Conklin, indicate that cleavage is an accurate means of separating 
materials already localized. My experiments on the eight-celled 
stage of Cerebratiihis show that the first localization of materials 
is in an apical-basal (polar) direction. It is probable, therefore, 
that at the four-cell and even at the two-cell stage this same ten- 
dency is the predominant one, so that at the four-cell stage there 
are four equivalent parts and at the two-cell stage two equivalent 
parts. However, in each of these two parts (taking the two-cell 
stage as an example) there is an apical-basal differentiation. A 
separation of a blastomere at this stage causes at first a half cleav- 
age, but the materials retain a relation to each other very similar 
to the normal whole relation as far as the apical-basal axis is con- 
cerned along which differentiation is assumed; the embryo, there- 
fore, can readily adjust itself to form a whole larva, having all 
the necessary materials present in the proper relations. 

In the fragment of the unsegmented egg this is not true. Here, 
according to all indications, there is a great activity in the mate- 
rials of the egg. If the egg is cut at an early stage (as in the un- 
fertilized egg) there is yet a considerable period of activity and 
movement of materials through which the egg must pass before 
the first cleavage takes place ; and, therefore, on the one hand a 
whole cleavage results, and on the other a normal whole larva 
is formed. Such is not the case, however, if we take the egg for 
instance after maturation not long before the cleavage. The egg 
is nearly ready for the first cleavage, the materials are arranging 
themselves for an equal distribution and the proper physical ten- 
sions for such a division are present. The egg is now cut and a 
portion of it removed. The cleavage ensues very quickly, for the 
physical machinery was already starting to act when the cut was 
made. The different materials are not separated in a precise way, 
even if the cut is vertical, for the cytoplasm is semi-liquid, and 
the materials, especially along the injured side, come into abnor- 
mal relations with each other, which cannot be regulated as in 

326 Charles Zeletiy. 

the earlier stages, because of the lack of the opportunity which, 
as stated above, is afforded the earlier ones {c. f. again the ex- 
periments on the frog's egg). The defects in the larvae have a 
definite relation to the position of the removed part of the egg 
because the disturbance of the protoplasm is greatest in the region 
near the cut. The already differentiated materials may thus be 
separated in an abnormal relation to each other and become un- 
naturally grouped by the cell walls of the ensuing divisions. In 
this manner, on the one hand the abnormality of the resulting 
cleavages, and on the other the defects in the larvze developed 
from the fragments, may be explained. The first normal cleav- 
age, however, divides the cell into two similar parts, each of 
which retains a relation between its differentiated materials very 
much like that of the whole egg. The conclusion is therefore 
reached that the relations of the materials in the isolated blasto- 
mere of the two-cell stage are more normal (/. e., more like those 
of the whole egg) than are those of fragments of the two-polar- 
body stage, and therefore the capacity of regulation to form a 
whole larva is greater in the former than in the latter. The 
mechanism of division which is disturbed by the cut in the un- 
segmented egg, and is capable of regulation if the cut is early 
but is disturbed if the cut is late, is also not disturbed in the iso- 
lated blastomeres of the two-cell stage, the cell division goes on 
as if the other blastomere were present, and a partial (one-half) 
cleavage results. 

It therefore seems probable that while in normal development 
cleavage is an aid in differentiation, in development after re- 
moval of a portion of the unsegmented egg (or segmented egg) 
it is a distinct detriment in so far as the attainment of the normal 
relations of the parts is concerned. For while on the one hand it 
isolates materials and allows a more accurate differentiation, on 
the other it restricts the power of regulation of the organism.^ 

^Yatsu has hinted at an explanation somewhat similar to the above. He 
says that the differences between the lai'vse derived from fragments of the un- 
segmented egg and those from isolated blastomeres of the two-cell stage may 
be due to dififerences in the accuracy of the separation of the materials in the 
two cases. 

Localization of the N emertine Egg. 327 

While the larvae developed from Isolated blastomeres of the 
two and four-celled stage show certain organic differences from 
the normal whole larvse, there is no indication in them of a specific 
local defect. The first trace of such a defect is reached in the 
eight-cell stage. Here, while larvse developed from lateral four- 
cell groups containing two cells of the upper quartet and two cells 
of the lower show the characters of a normal larva (except for 
asymmetry in arrangement) larvae from the upper quartet al- 
ways possess an apical organ and lack an enteron, and those from 
the lower possess an enteron and lack an apical organ. There is 
thus a very distinct differentiation along the apical-basal (polar) 
axis. It is an interesting fact that this first distinctive morpho- 
genic localization is coincident with the first inequality in cleavage, 
the Inequality being In the same direction. Projecting backward 
this differentiation — that Is, assuming that an apical-basal differ- 
entiation has been going on for some time before the eight-cell 
stage — naturally there would be no indication of It In the Isolated 
blastomeres of the two or four-cell stages because the cleavages 
are vertical. Likewise there may be a similar differentiation in 
the unsegmented egg; for, while my results on cleavage defects 
cannot be analyzed as showing any specific relation to the indi- 
vidual kinds of egg defects, the observations of Yatsu on morpho- 
genlc defects do show such a relation. 

The experiments on the eggs of Cerebratuhis marginatits, to- 
gether with the former ones on C. lacteus, seem therefore to indi- 
cate that at the eight-cell stage the formative materials of the 
egg are definitely localized in an apical-basal direction, and the 
experiments of Yatsu on morphogenic defects In larvae resulting 
from unsegmented eggs of the later maturation stages show a 
similar apical-basal differentiation. That this process of apical- 
basal differentiation is a progressive one In the unsegmented egg 
Is Indicated by the whole character of the cleavage in fragments 
of unfertilized eggs, and by the progressive departure from this 
character up to the first cleavage, and the corresponding increase 
In defects of larvae developed from the fragments, though In the 
latter case the continuity of the result seems to be masked by the 
development of whole larvae from isolated one-half and one-fourth 

328 Charles Zeleny. 

blastomeres. An explanation of this has, however, been offered. 
The first two cleavages being perfect apical-basal ones, the iso- 
lated blastomeres cannot be expected to show other than perfect 
larvae, assuming only a slight regulation overcoming lateral 
asymmetry, notwithstanding the partial cleavage, which after all 
is only quantitatively partial (=^ or 34 of a pattern). At the 
same time the fragments of unsegmented eggs can never be said 
to contain the materials divided accurately with respect to an 
apical-basal axis because, in the first place, the cut is never per- 
fectly vertical, and in the second place, the rounding in of the 
edges after such a cut causes a disarrangement of the materials 
which must result in unequal distribution at the first cleavage. The 
ability to regulate such an unequal distribution must of course 
largely depend upon its extent and character. The greater op- 
portunity given to fragments of unfertilized eggs to regulate such 
differences in distribution (if any) before cleavage takes place 
may to some extent explain the whole character of the cleavage 
in such fragments without the assumption of a perfectly Isotropic 
egg, an assumption which is contradicted by the evident polarity 
of the egg at this period as indicated by the eccentricity of position 
of the nucleus and the presence of the basal protuberance. The 
existence of such an apical-basal differentiation in the unsegmented 
egg was indeed already indicated by Professor Wilson's result on 
certain eggs in which the basal portion was removed by a hori- 
zontal cut and which showed the basal quartet of the resulting 
eight-cell stage much smaller in comparison with the upper than 
In the normal whole egg. 

The data on localization of formative factors in the egg be- 
fore cleavage and during the early segmentation stages may, 
therefore, be provisionally stated In the following form for Cere- 
bratuliis, If an Intimate relation is assumed between localization 
of visible materials and localization of formative factors. The 
unfertilized egg before the beginning of maturation already shows 
evidences of a polarization which necessitates the assumption of 
a heterogeneity In material. Upon this basis, and In the same 
apical-basal direction, later differentiation proceeds. 

During the preliminary maturation stages and after fertlliza- 

Localization of the Nemertine Egg. 329 

tion there are profound changes in the distribution of materials 
in the egg, and these changes seem to be accompanied by an in- 
creased apical-basal differentiation. The first two cleavages being 
vertical and equal merely effect a quantitative and not a qualita- 
tive separation of materials, but the third plane of division, a 
horizontal one, bringing about an unequal division, separates the 
egg into two qualitatively different parts. That such is the case 
is absolutely demonstrated by my experiments on the eight-cell 
stage in which I obtained complete larvae from lateral four-cell 
groups, larvse with an apical organ but without an enteron from 
the upper quartet and larvae with an enteron but without an apical 
organ from the lower quartet. 

Hull Zoological Laboratory, University of Chicago. 
February 29, 1904. 



The so-called "polarity" exhibited in the regeneration of ani- 
mals has suggested the idea to a number of writers that the phe- 
nomenon might be related to, or the outcome of differences in 
potential in different regions; or, in other words, of electrical polar- 
ity. The term "polarity" itself, which has been generally adopted 
to express a sort of stereometric relation in the regeneration of 
living things, suggests in certain striking ways the polar relations 
observable in many electrical phenomena, and invites a direct com- 
parison between the two. 

The only experiments that have been undertaken to test directly 
this question are the recent ones by Mathews'^ on certain hydroids 
and on the tail of Fundulus. The interesting results reached by 
Mathews, while leaving the problem, so far as the main issue is 
concerned, still an open one, showed the importance of .further 
examination of the subject. Mathews avoids, it is true, making 
a direct comparison between physiological "polarity" and the 
polarity present in electrical phenomena, and speaks rather of the 
rate of growth of certain regions in comparison with others; but 
if there is in reality any fundamental relation between the phe- 
nomena in question, we should expect to find some expression of 
it in the polar, or, more generally, the stereometric relations of 
the parts. 

If, for instance, the development of a head at the anterior end 
of a piece and of a tail at the posterior end is connected with dif- 
ference of potential in the two regions, we might hope to get evi- 

iRlectrical Polarity in the Hydroids. A. P. Mathews. Am. Journ. Physi- 
ology, Vol. VIII, No. IV, Jan. i, 1903, pp. 294-299. 

332 T. H. Morgan and Abigail C. Dimon. 

dence of this by means of the galvanometer. If this relation 
should be found to exist, there is a further opportunity of test- 
ing the validity of the conclusion in the case of axial heteromor- 
phosis. For this reason we selected the earthworm for our study, 
since in the earthworm it had been shown by one of us that there 
regenerates from the anterior cut surface of posterior pieces not a 
head, but a tail. We should expect to find under these circum- 
stances a reversal of the potential in this region,* when compared 
with an anterior cut surface in the more anterior regions of the 
worm. The following pages give the results of our examination. 


Both Lumbricus terrestris and Allolobophora foetida were used. 
Since the results appeared to be similar for both species, Lum- 
bricus, being larger and showing on the whole greater differences 
of potential, was preferred when available. In all, sixty-four 
worms were used, the number of tests made upon any one varying 
from one to seventeen. Differences of potential were detected by 
a Rowland-d'Arsonval galvanometer, connected with a pair of 
non-polarizable electrodes. Since the regulation of these electrodes 
was found to be troublesome, and since they were liable to intro- 
duce a source of error into the readings, their manufacture and 
regulation had best be described. Two glass tubes about three 
inches long, somewhat smaller at one end, were plugged at the 
small end with kaolin or filter paper, moistened with normal 
(0.85%) salt solution. The plug extended about a half Inch be- 
yond the end of the glass tube. When made of kaolin, the end, 
after being used, could be easily broken off and re-formed from 
fresh material. When made of filter-paper, the tip, if it became 
contaminated from touching the worm, could be cut off. Above 
the plug each tube contained a saturated solution of zinc sulphate, 
into which projected through a cork a small amalgamated zinc 
electrode, connected by a wire with one of the poles of the gal- 
vanometer. Since the instrument was so sensitive that the slight- 
est loss of equilibrium was at once registered by a deflection of the 
mirror, great care had to be exercised in keeping the junctions of 
wires and zinc electrodes dry, and in balancing the other elements 

Physiological ''Polarity" and Electrical Polarity. 333 

of the electrodes. As the electrodes were found to deteriorate 
rapidly when used, it was necessary to examine them at frequent 
intervals, and when the deterioration became so great as to affect 
seriously the value of the readings on the worm, to readjust the 
parts. This could be accomplished in several ways, of which some- 
times one and sometimes another was most effective. A fresh 
zinc electrode could be substituted, the zinc electrodes could be 
freshly amalgamated, or they could be polished by rubbing them 
with sandpaper. The current could sometimes be affected by 
moving one zinc so that a greater or less surface was immersed 
in the zinc sulphate. Putting fresh zinc sulphate solution into 
the tube nearly always produced a distinct effect. All these means 
of regulation proved more or less temporary, and often, when 
they all failed to balance the electrodes, fresh ones had to 
be made. The electrodes in which filter-paper was used proved 
much more constant and easy to regulate than the clay ones, and 
if they were often renewed, it seemed safe to employ them. 

A positive deflection of the galvanometer meant that the poten- 
tial at the right hand electrode was higher than that at the left 
hand, i. e., the current through the galvanometer flowed from 
right to left. Since in nearly all cases the left hand electrode was 
placed on the worm anterior to the right hand one, a positi\"e de- 
flection meant that the current flowed through the galvanometer 
from the posterior to the anterior electrode. 

The electrodes were applied, sometimes both to the dorsal sur- 
face of the worm, sometimes to the ends made by transverse sec- 
tions through the body, and sometimes one to the surface, and the 
other to a cut end. The results will be considered according to 
the position of the electrodes. 



It will be seen from the records given in the following selected 
tables that however general certain results appear to be, never- 
theless some individuals show irregularities. This uncertainty in 
the results can be in part at least accounted for by the following 
considerations. Secretions in different regions of the body, the 

334 T. H. Morgan and Abigail C. Dimcn. 

flow of blood from cut surfaces, the excretion of slime from the 
skin, local or general muscular contractions might all tend to af- 
fect the results. Such effects could sometimes be distinctly seen 
in an increase or decrease in the deflection of the galvanometer. 
It seemed possible that the presence or absence of food in the 
digestive tract might in itself or by stimulating the flow of digest- 
ive fluids influence the distribution of electric potential, but this 
was not observable from tests made with worms that had been 
starved for several days. 

Experiment i. Lumbricus terrestris starved two days.^ Cut 
ends anterior at several levels. The left hand electrode was ap- 
plied to the cross-section and the right hand electrode to the dor- 
sal surface one-third to one-half inch behind the section. The 
zero point of the galvanometer was 27.1, and the effect of the 
electrodes varied, deflecting it between the limits 26.1 and 29.3, 
both of which deflections are less than those caused by the worm 
itself, and may therefore be disregarded. The galvanometer 
readings at different levels on the worm were as follows : 

Cut at fourth segment. ... a. 45.0+ (off scale) Surface positive 

Cut at fourteenth segment. b. 35.5 

Successive sections between 1 c. 45.0+ (off scale) 

the fourteenth segment J d. 42.5 

and the middle of the 1 e. 40.0 

worm I r. 43.5 

Successive sections posterior I g. 11.5 Cut end positive 

to the middle J h. 34.0 Surface positive 

[ 1- 32.2 

The readings are all definite and represent the state of affairs 
in a majority of the worms examined. From these data it will 
be seen that when a worm is cut in two it is found that in the 
anterior regions of the worm the anterior cut end of the posterior 
piece is negative with respect to the near-lying surface. In the 
posterior regions of the worm where there was more variation the 
differences in potential were usually less, and sometimes reversed in 

^Since the results seemed not to be affected by the absence of food from 
the digestive tract, this specimen was chosen because a more complete series 
of sections were made from it than from any unstarved worm. 

Physiological "Polarity" and Electrical Polarity. 335 

direction. Irregularities were often observed at about the fif- 
teenth segment, where the male reproductive organs open to the 
exterior and the crop-gizzard region begins. Irregularities also 
occurred at the girdle. In the worm used for this experiment 
negative deflection of the galvanometer occurred only once, im- 
mediately behind the middle of the worm, but the positive deflec- 
tion was diminished at the fourteenth segment and at the other 
cut ends in the posterior regions of the worm. 

Experiment 2. Lumbricus terrestris. Cut end posterior. The 
zero point of the galvanometer was 27.3, and the electrodes 
caused a deflection to 27.5 at the beginning and to 30.0 at the end 
of the experiment, which does not affect the sense of the deflection 
for any reading. The galvanometer record for the different levels 
is as follows : 

[a. 28.0 Cut end positive 
Successive sections from the posteriori L. 19.5 Surface positive 

end to the middle of the worm. . . c. 23.0 

I d. 23.5 ;; ;; 

Section about the middle e. 24.0 

Section anterior to middle f. 25.0 

Section at 23rd segment g. 39.0 Cut end positive 

Section at 15th segment h. 20.0 Surface positive 

Section anterior to 15th segment. ... i. 17.7 " 

The surface has a higher potential than the cut end at all but 
two levels, one where only a few segments are cut oflf from the pos- 
terior end, where the difference of potential is small; and the other 
at the twenty-third segment. The reversal of current at the extreme 
posterior end occurred in all the worms in which a full series of 
sections was made. The twenty-third segment is between the fif- 
teenth segment and the girdle. In this region two other worms 
showed possible cases of reversal of current, though the more usual 
condition was that the current was reversed at the fifteenth seg- 
ment or at the girdle, or at both points, but between them flowed 
in the same direction as in the rest of the worm. The worm used 
in this experiment showed the same distribution of potential that 
is found in a majority of individuals, except at the fifteenth seg- 
ment, where one-half of the worms tested showed a reversal of 

336 T. H. Morgan and Abigail C. Dimon. 

current, the other half agreeing with this experiment in showing 
no reversal. 

From experiments i and 2 it may be assumed that when an 
earthworm is cut in two, the transverse section commonly repre- 
sents a point of lower potential than the uninjured surface near to 
it. At the cut end chemical changes no doubt take place as a re- 
sult of the fluids there set free, and of the general breaking down 
of tissues. These conditions might be expected to alter the elec- 
trical potential at the cut end, and presumably, where these changes 
are greatest, the alteration of potential will be greatest. Since, 
however, the transverse section is usually lower in potential than 
the uninjured surface near it, whether the section be at the anterior 
or posterior end of the piece, the difference in potential cannot bear 
any relation to the kind of regeneration that is to take place. 

Having illustrated the more regular and more usual conditions 
of distribution of potential between surface and cross section, two 
illustrations of the exceptions to this condition, sometimes met 
with, will now be given. In the first a distribution of potential 
different from the average occurred throughout the whole worm. 
In the second, a short piece of a worm showed unusual conditions. 

Experiment 3. Lumbricus terrestris, young. Cut end posterior. 
The zero point of the galvanometer was 27.4, and the electrodes 
deflected it to 26.3 at the beginning of this experiment, and to 28.9 
at the beginning of the next experiment made that day. Some 
doubt may, therefore, be thrown on two of the readings given 
below — those at the girdle and at the fifteenth segment. The direc- 
tion of deflection in these cases is probably correct, but the amount 
is small. The readings were as follows :- 

Posterior to middle a 24.0 Surface positive 

Just back of girdle b 29.0 " " 

A .^ • ^ • J, ( c 33. c Cut end positive 

Anterior to girdle 3 -^-^ -i ^ ^^ 

] d 34.6 

At 15th segment. e 26.5 Surface positive 

I f 32.8 Cut end positive 

Anterior to icth segment ' ? ^o-O 

^ ' h 34.9 

i 35-0 

Physiological "Polarity" and Electrical Polarity. 337 

Here we see the cut end with a higher potential than the surface, 
except at the girdle and in front of the fifteenth segment, where 
the current is reversed. The worm was small and immature, but 
as other young worms gave the same sort of readings as the ma- 
jority of mature worms, the peculiar results cannot be due to im- 
maturity. In fact, it is difficult even to guess why this worm should 
give such different responses from the others. Another case 
was also recorded in which the cross section was anterior and of 
higher potential than the neighboring surface for a series of seven 

Experiment 4. Lumbricus terrestris. Piece one inch long from 
near the posterior end of the worm, with a cut surface at each end 
of the piece. Zero point of galvanometer, 28.6; electrodes, 30.0. 

Electrodes at anterior end and 

middle of piece 26.3 Anterior cut end positive 

Electrodes at posterior end and 

middle of piece 27.1 Surface positive 

Electrodes at both ends 27.0 Anterior cut end p^ositive 

In this experiment the anterior electrode is positive with respect 
to the posterior one, whether it be on an end or on a surface, and 
we get a constant direction of current from before backward. The 
worm from which the piece was cut gave the usual results for 
the other readings made on it. The piece was cut out by two 
consecutive cuts with no appreciable time between them, so that 
the freshness of the cut could not, as it does in other cases given 
later, influence the result. One other case resembled this one, 
while two cases showed conditions in which the middle was posi- 
tive with respect to both ends, and three cases showed conditions 
in which the middle was negative with respect to both ends. 

It has been stated that marked changes in the distribution of 
potenial between the surface and cross section often occur at the 
fifteenth segment and at the girdle region. When a worm is cut 
in two at the fifteenth segment the cut end has usually a higher 
potential than the surface either anterior or posterior to it, or when 
short pieces of a worm are cut with one end at the fifteenth seg- 
ment, that end is positive to the other, whether it be an anterior 

338 T. H. Morgan and Abigail C. Dimon. 

or a posterior end of the piece. When, however, a worm is cut 
in two immediately anterior or immediately posterior to the girdle, 
the girdle has a lower potential than either cut end. Both these 
regions, therefore, show a state of affairs different from that in 
other parts of the worm. When worms were cut in two at a 
series of points it was found that at the fifteenth segment there was 
a change in direction of the current in twelve cases out of nine- 
teen, and at the girdle in eight cases out of fourteen, four cases 
in which the change was not very pronounced being included in the 
first series, and two in the second. The change in direction of cur- 
rent, though by no means uniform, is rather more likely to occur 
than not, and may perhaps be connected with substances secreted 
by the organs at these levels. 

If the distribution of potential in the earthworm resembles the 
distribution of potential in a resting muscle, we might expect that 
the difference of potential between the electrodes would vary ac- 
cording to the position of the electrodes. A series of experiments 
was tried, in which one electrode was kept stationary at a trans- 
verse section, and the other moved along to different positions — 
usually one near, one half way between the ends, and one on the 
skin at the end opposite the transverse section. Great variation 
in the deflection of the galvanometer was always observed for 
these different positions, but it was by no means regular. The 
most common case was that the deflection- was greatest when the 
electrodes were near one another, decreasing as they moved away, 
and sometimes even changing to an opposite direction when they 
were at opposite ends of the worm. In this series of experiments 
we have not only the complicating conditions already mentioned, 
but also the factor of resistance which would be approximately 
proportional to the distance between the electrodes, and if appre- 
ciable would modify the results in the way stated. The problem 
is one of greater complexity than that of the distribution of poten- 
tial in the comparatively homogeneous tissue of the muscle, where 
the resistance is small. If the earthworm were homogeneous as 
regards electrical conductivity, and a difference of potential were 
set up by means of a transverse section, the point of lowest poten- 
tial would be in the middle of the cut end, and of highest potential 

Physiological "Polarity" and Electrical Polarity. 339 

at the opposite end of the worm. If, however, resistance varied 
in different parts of an unhomogeneous tissue, the difference of 
potential observed between two points would be a resultant be- 
tween a tendency to a regular rise of potential and an irregular 
distribution of resistance, and the recorded distribution of poten- 
tial would, therefore, be irregular. In point of fact, however, as 
mentioned above, other causes of irregularity may be added to 
those due to resistance, and many arrangements of potential were 
observed in the seventeen cases tested. 


Experiment 5. Lumbricus terrestris. The zero point of the 
galvanometer was 28.6, and the electrodes deflected it to 27.9 or 
27.8, in an opposite direction from deflections caused by the worm. 
This experiment was made in order to see what the electrical con- 
ditions are on the surface of the worm. Three readings were 
taken, as follows : 

a. Earthworm cut in two at the mid- 

Posterior piece. One electrode on 
the surface at the section and the 
other on the surface, posterior to 
the section, but near It 30.3 Posterior positive 

b. Same worm. Short piece cut from 

posterior half of animal. 
One electrode on the surface at the 

anterior section, and one on the 

surface posterior but near 33.9 " " 

One electrode on the surface at the 
anterior section, and one on the 

surface at the posterior section. 35.0 " " 

The direction of current was the same as would be expected if 
one electrode were applied directly to the cut end, and the other to 
the uninjured surface near the end. On the assumption that at 
any level the conditions of the surface fairly represent those in the 
interior of the worm, by testing the distribution of potential at 
the surface of an uninjured specimen It may be possible to get some 

340 T. H. Morgan and Abigail C. Dimon. 

idea of the distribution M'ithin the worm. This was done in the 
following experiment. 

Experiment 6. Lumbricus terrestris. Both electrodes applied 
to the surface. The zero point of the galvanometer was 26.9, 
and the electrodes varied from 26.7 to 22.0 after the third read- 
ing, when they were regulated to 24.3, and at the end registered 
only 26.0. The readings of the galvanometer were as follows: 

One electrode at anterior end, the other 

near 32.0 Head end negative 

One electrode at anterior end, the other 

at middle 30.0, then up, off scale. 

Head end negative 
One electrode at anterior end, the other 

at posterior end 16.0 " " positive 

One electrode at posterior end, the 

other at middle 28.0 Tail end positive 

One electrode at posterior end, the 

other near i8.oi ,, ,, 



It is difficult to explain these data so that they are consistent. 
There Is no one point that has a high potential relative to all 
others, though the deflections are sufficiently strong to indicate 
that they are not due to variability In the electrodes themselves. 
The variation may perhaps be due partly to local muscular con- 
traction and partly to the excretion of slime at points on the sur- 
face, for in Allohophora, where the body cavity fluid extruded 
through the dorsal pores is yellow and noticeable. Its excretion 
was observed to have a great effect upon the deflection of the 

Different worms, too, show the greatest differences as to their 
reactions when electrodes are touched to different parts of their 
surface. In general, the two ends tend to have a lower potential 
than other parts of the surface, and the middle tends to have a 
higher potential with respect to points on either side of It. At 
the girdle and at the fifteenth segment, however, the results are 
more definite, as Is shown in the two following experiments. 

Physiological "Polarity" and Electrical Polarity. 341 

Experiment J. Lumbricus terrestrls. Zero point of galvanom- 
eter, 28.6; deflection caused by electrodes to 30.5. 

One electrode at girdle, the other anterior 

to it, and near 20.5 Girdle negative 

One electrode at girdle, the other posterior 

to it, and near 39.0 " " 

Experiment 8. Lumbricus terrestris. Zero point of galvanom- 
eter, 28.6; deflection caused by electrodes to 30.0, at end of 

One electrode at 15th seg- 
ment, the other ante 
rior to it and near. . 29. S 15th segment negative, probably 

One electrode at 15th seg- 
ment, the other poste- 
rior to it and near. . . .31.5 " " " 

The girdle is definitely of a lower potential than the surface 
near it, anterior or posterior, and this was found to be the case 
for four worms tested. At the fifteenth segment the difference 
was not so great, and though this region was negative with respect 
to a surface posterior to it, with respect to one anterior it was only 
very slightly, or, perhaps, not at all so. In another worm the fif- 
teenth segment was evidently positive with respect to a surface an- 
terior to it. 


Experiment 9. Lumbricus terrestris. Pieces cut out from 
worm. The zero point of the galvanometer was 27.3. The 
electrodes deflected it to varying amounts. In two cases, namely 
the fourth and the seventh readings in the table, where the elec- 
trodes deflected the galvanometer to 26.1 and to 29.5, respectively, 
these deflections come near the readings given by the worm. In 
the other cases it is not necessary to take deflection caused by the 
electrodes into account, since they would not affect the direction of 
the reading. The data are as follows : 

342 T, H. Morgan and Abigail C. Dimon. 

(a) Anterior half of worm, both ends cut 

(long piece) 16.6 Anterior positive 

(b) Short piece from anterior part of an- 

terior half of worm, the posterior 
end cut somewhat later than the 
anterior end 22.3 

(c) Same. Anterior end freshly cut. . . . 32.5 Posterior positive 

(d) Short piece from middle part of an- 

terior half of worm, posterior end 
more freshly cut 27.1 



(e) Same piece, anterior end freshly cut. 36.2 " " 

( f ) Short piece from anterior part of pos- 

terior half of worm 33-0 " 

(g) Short piece from middle of posterior [ 28.5 Anterior end 

half of worm \ 24.0 positive probably 

(h) Anterior half of (g), posterior end 

freshly cut 22.0 Anterior positive 

(i) Posterior half of (g), anterior end 

freshly cut 34-0 Posterior positive 

In this worm, when the two ends were cut at approximately 
the same time, which happened in (a), (f) and (g), the piece 
from the anterior half had its anterior end positive, and the two 
pieces from the posterior half had their posterior ends positive. 
In the majority of worms tested, when the two ends of a piece 
were cut at the same time, the anterior end was positive, regard- 
less of the position of the piece on the worm.^ If, however, the 
two ends were cut at different times, which in this worm occurred 
in six pieces, the end which' had been cut most recently generally 
had a lower potential than the other. Since the testing with the 
electrodes on a transverse section and an uninjured surface near 
that section the end was usually found to be at a lower potential 
than the surface, the fall of potential was supposed to be due to 

^When, however, the pieces were long (one-half the worm or more), in a 
majority of cases the posterior end was positive with respect to the anterior. 

Physiological "Polarity" and Electrical Polarity. 343 

changes accompanied by the escape of body fluids at the cut end. 
Since, when the electrodes are on two transverse sections, the one 
that is more recently cut is of a lower potential than the other, it 
would appear that the causes that determine the fall of potential 
are such as decrease in the course of a short time. The fluids 
that escape from the cut end dry rather rapidly, whereas the tissue 
cells, breaking down, are not built up for several days. The first 
fall of potential, then, is probably largely due to escape of blood 
or other fluids from the section, while the slighter permanent effect 
may be due to the breaking down of tissue cells. When there 
was but a short time between the two cuts the lower potential did 
not always occur at the fresher one, which may have been partly 
because fresh fluids were still coming from the earlier section. In 
the anterior half of the worm, also, there was great irregularity, 
which may be partly due to difi^erent digestive fluids at different 
regions of the digestive tract producing varying electrical activi- 

The differences of potential between two transverse sections are 
probably a resultant of the factors that cause difference of poten- 
tial between a section and a surface. If a piece be cut out from 
an earthworm by two transverse sections, there will be a difference 
of potential between each end and the surface between the ends. 
If the differences of potential between each end and the surface 
are equal and opposite they will balance one another, and there 
will be no difference of potential when electrodes are applied to 
the two ends. If, however, the differences of potential between 
the two ends and the surface are unequal, their resultant will de- 
termine a difference of potential between the two ends. This is 
illustrated by experiment 10. 

Experiment 10. Lumbricus terrestris. Short piece. The zero 
point of the galvanometer was 28.6, and the electrodes registered 
slightly above this at the beginning of the experiment. The read- 
ings were as follows : 

(a) One electrode applied to anterior end, the other to the 

middle of the piece 25.0 

(b) One electrode applied to posterior end, the other to the 

middle of the piece 30.8 

344 T. H. Morgan and Abigail C. Dimon. 

(c) One electrode applied to one end, the other to the other 

end of the piece 28.0 

If we disregard the deflections caused by the electrodes, the 
following are the departures from the normal: (a.) == — 3.6, 
(b.) = +2.2, (c) = — 0.6. Theoretically, if (c.) were the re- 
sultant of (a.) and (b.), it would equal — 1.4, but considering 
how variable the conditions were always found to be, — 0.6 pre- 
sents a fairly close agreement. In all other cases but one, where 
similar tests were made, the results agreed in like manner with the 


In addition to the readings made from worms that had been 
freshly cut in two, a series of readings were made on worms in 
which the process of regeneration had proceeded for a number 
of days. The regenerating worms were divided into four groups : 
( I ) those in which a few anterior segments had been cut off and 
regeneration was taking place at the anterior end of the long 
piece; (2) those in which the worm had been cut in two in the 
middle and regeneration was taking place at the posterior end of 
the anterior piece; (3) those in which the worm had been cut in 
two in the middle and regeneration was taking place at the an- 
terior end of the posterior piece; (4) those in which the worm had 
been cut in two at the fifteenth segment and regeneration was tak- 
ing place at the anterior end of the posterior piece. They were 
allowed to regenerate from twenty-five to thirty-two days, and in 
the course of that time five or six tests were made on most of 
them at intervals of a few days. One electrode was applied to the 
regenerating tip, and the other to the old surface a short distance 
from the tip. In all one hundred and fourteen readings were re- 
corded, the results of which may be summarized as follows : 

Group (i) 31 cases, end positive ; 9 cases, end negative 

Group (2) 20 " " " 18 " 

Group (3) 15 II 

Group (4) 8 " " " 2 " 

Total 74 " " " 40 " 

Physiological "Polarity" and Electrical Polarity. 345 

If we regard only the readings made when regeneration had 
proceeded more than twenty-one days, in thirty-six cases the end 
was positive with respect to the surface, and in eleven cases it was 
negative, a much more definite result than when all cases are con- 

The conditions in a regenerating tip do not, therefore, agree 
with those at a freshly cut end, for the current flows in an oppo- 
site direction. To be sure, the processes occurring during regen- 
eration are not the same as those occurring during the breaking 
down of tissues, and the latter may be the predominant ones im- 
mediately after a cut is made. The subject, however, needs fur- 
ther investigation before the causes for this reversal can be more 
than surmised. 


If we look for a relation between electrical polarity in the worm 
and rate of regeneration, as Mathews has suggested, we find it 
as difficult to demonstrate as the difference between electrical and 
physiological polarity. If the average deflection of the galvanom- 
eter was greater at certain levels where regeneration is known 
to be rapid than at other levels where it is slow, the connection 
would be established. For instance, the regeneration of a tail at 
the posterior end of a worm when only a few posterior segments 
are cut off, is exceedingly rapid, whereas the regeneration of a 
heteromorphic tail at the middle of a worm is very slow. The 
average deflection of the galvanometer in the former case for 
three readings is, however, 3.4, with the cut end positive instead of 
negative in every case. In the latter case the average deflection 
for seven readings is 3.6 (with the end negative) , with no extreme 
readings to brtng it up. The region in the middle of the worm, 
where a tail is to regenerate from the posterior end of the anterior 
piece, gives an a^^erage deflection of 4.7 for seven readings from 
different worms. When five or six segments are cut from the an- 
terior end of the worm the average deflection for seven cases was 
4.6 at the anterior end of the long piece. At the posterior end of 
the short piece, regeneration would be very slow, and at this end 

346 T. H. Morgan and Abigail C. Dimon. 

only two readings were made, one giving a very slight result, the 
other deflecting the galvanometer to 5.2. 

If we attack this subject by another method, namely, by mak- 
ing a direct comparison between two freshly-cut ends on one worm, 
the results are equally indefinite. When five or six segments were 
cut from each end of the worms, of a series of fifteen readings on 
different worms made with one electrode at the anterior-cut end 
and the other at the posterior-cut end, the posterior end was posi- 
tive in eleven and the anterior end positive in four. When the 
worm was cut through the middle and at the posterior end, the 
posterior end was positive in three cases, and the middle in one 
case. When the cuts were made through the middle and anterior 
end, in five cases the middle was positive, and in two the anterior 
end positive. From these considerations it would therefore ap- 
pear that no invariable connection between rate of regeneration 
and electrical polarity exists in the earthworm, at least as measured 
on a freshly cut surface. 

From the foregoing experiments we conclude : 

( 1 ) That a freshly cut end of an earthworm is generally nega- 
tive with respect to a near-lying uninjured surface. 

(2) That the freshness of the cut surface has an Important in- 
fluence In determining the amount of difference of potential. 

(3) That the result is often complicated by the presence of se- 
cretions or exudations on the surface, or by the presence of certain 
organs at the cut end, or by the contractions of the worm, etc. 

(4) That in the region of the girdle and also in the region of 
the fifteenth segment (near which the crop and gizzard lie), the 
results are often different from those elsewhere. 

(5) That there Is no apparent relation between the differences 
In potential at freshly cut surfaces and the kind of regeneration 
(head or tail) that occurs. 

(6) That cut surfaces from which heteromorphic growth 
would take place show the same sort of differences in potential 
as those from which orthomorphic regeneration occurs. 

(7) That the differences in potential present when a cut sur- 
face Is exposed can probably be accounted for by the chemical 
changes taking place at the surface; and these need have, and do 

Physiological "Polarity" and Electrical Polarity. 347 

not appear to have, any relation to the kind of regeneration that 
takes place. 

(8) That when, on the other hand, the cut surface is allowed 
to heal, and when later a new structure has begun to appear, the 
differences in potential between the new and the old parts {as 
measured on the surface only) are not such as can be made to ac- 
count for the difference in the kind of part (head or tail) that is 
regenerating. Here also many complications enter into the re- 
sult and make it difficult to draw satisfactory conclusions. 

(9) No definite relation was found between the rate of growth 
and the fall of potential between an uninjured surface and a cut 



In a paper by Professor Morgan^ an account was given of an- 
terior regeneration from three different levels in earthworms. 
The results seemed to show that the internal factor determining 
the formation of a heteromorphic tail might be the presence of 
the stomach-intestine at the regenerating surface, and at Professor 
Morgan's suggestion and under his direction, the following ex- 
periments were undertaken. An attempt was made to test this 
view by means of more exactly localized sections made near the 
level of the beginning of the stomach-intestine. 

In Allolohophora foetida, the species used, the oesophagus ex- 
tends to the fifteenth segment, the crop lies in the fifteenth and 
sixteenth, the gizzard in the seventeenth and eighteenth, and at 
the nineteenth begins the stomach-intestine, which extends pos- 
terJorally through the rest of the worm. The external openings 
of the vasa deferentia on the fifteenth segment served as con- 
venient landmarks for determining the level of the section. The 
worm was cut in two, the short anterior piece dropped into alcohol, 
and its number of segments counted so that the exact level of the 
cut could be recorded. The posterior piece was then left to re- 
generate from forty-eight to one hundred and twenty days, when 
it was killed, and sections made for study. In some cases a new 
head, and other cases a new tail regenerated from the anterior end 
of the posterior piece. In the regenerating head the new stomo- 
daeum usually did not open into the old digestive tract, which 
closed anteriorly, and no definite pharynx formed, A dorsal brain, 
connected with the ventral nerve cord was usually present. Since 
these conditions represented the most usual form of head regenera- 

lExperimental Studies of the Internal Factors of Regeneration in the Earth- 
worm. Arch, fiir Entwickelungsmech. der Organ. Bd. XIV. pp. 562-591. 


Abigail C. Dimon. 

tion at the levels of these experiments, the cases in which they ex- 
ist are classified in the table as a separate group under Head A. 
Cases where the brain lies anterior and even ventral to the level 
of the digestive tract; where the nerve cord ends without form- 
ing a brain; or where there is no mouth invagination, though the 
brain is well developed, are classified as Head B. The heads of 
group B look less like a normal head than those of group A, and 
yet are very evidently to be classified as heads rather than as tails. 
The distinctive features indicating a heteromorphic tail are the 
formation of a number of segments, the opening of the digestive 
tract to the exterior through a new anus, and the ending of the 
nerve cord ventrally, without a brain. Tails possessing these fea- 
tures are put in group A, while those in which any of these features 
are absent, are put in group B/ 

In all, one hundred and seventeen worms were examined, with 
the results given in the table. The number of the segment given 
at the head of each column locates the level at which the worm was 
cut in two, and both the actual number of worms and the percent- 
ages are given under each class. 


13 h 
















Back of 

1 *h 















Head, A.. 
Head, B.. 












1 , 










Tail, B... 


Tail, A. . . 










Total . . 














iJn only twenty-four cases out of one hundred and seventeen did the old di- 
gestive tract open to the exterior through the new mouth or anus. This occurred 
ten times in head regeneration, ten times in heteromorphic tail regeneration, 
and four times in cases classified as uncei^tain. Seven of the twenty-four cases 
occurred when the worm was cut in two in front of the sixteenth segment, and 
the other seventeen when it was cut behind the eighteenth segment. 

Heteromorphic Tail in Allolohophora. 351 

Since there were but few worms cut further back than the eigh- 
teenth segment, and since the stomach-intestine begins at this level, 
all the observations made on worms cut posteriorly to the eigh- 
teenth segment were brought into one class. It is worth noting, 
however, that of the four heads regenerating at these levels, three 
formed from worms cut at the nineteenth segment, while the 
fourth formed from a worm in which the exact level of the cut 
was not noted. The percentages In the different classes, though 
based on a small number of cases, yet bring out clearly one or two 
points. When a worm was cut in two In front of the stomach- 
intestine, In no case was a heteromorphic tail formed. The per- 
centage of cases In which a head was formed grows less as the sec- 
tion Is made further back on the worm, the fall of percentage 
being very great Immediately behind the gizzard. This tends to 
support the hypothesis that the formation of a heteromorphic tail 
is favored by the presence of the stomach-intestine near the cut 
end, though when the section is not more than one segment back 
of the gizzard a head is sometimes formed. 

Though the preceding experiments seem to show that the devel- 
opment of a heteromorphic tail is connected with internal struc- 
tures in the worm, they leave untouched the question of the kind 
of regeneration that takes place from posterior ends of anterior 
pieces cut anterior to the stomach-intestine. This point should be 
determined, and I hope In the future to undertake a set of experi- 
ments in which the posterior regeneration from anterior pieces 
will be observed. 





On the surface of the coral reefs guarding the harbor of Apia 
(Samoa) the five-rayed starfish, Linckia pacific^, with its long, 
slender, smooth, sky-blue arms, is the most conspicuous and abun- 
dant echinoderm in a place where echinoderms abound. Associ- 
ated with it, and similarly blue and conspicuous, although smaller, 

Fig. I. Linckia diplax, regener- 
ating from a single arm; note these 
new arms and new disc with madre- 

Fig. 2. Linckia dip/ax, re- 
generating from a single arm; 
note four new arms and new 
disc with madreporites. 

is the species L. diplax. Both for number of species and wealth 
of individuals, the Apia reefs are distinguished by their star- 
fish, sea-urchin and holothurian fauna. In collecting on these 
reefs during several weeks in the summer of 1902, as a member of 


Vernon L. Kellogg. 

the U. S. Bureau of Fisheries' Samoan Explorations party, my 
attention was particularly attracted by the many examples of star- 
fishes with regenerating arms, and I gave some special care to pick- 
ing up such specimens. From this material the figures here pre- 
sented have been drawn and in themselves tell how effectively this 
capacity for restorative regeneration obtains in this species. 

Morgan calls attention in his "Regeneration" (1901, p. 102 
and elsewhere) to the assertions of some authors that starfishes 

Fig. 3. Linckia dip lax, 
regenerating from a sin- 
gle arm. 

Fig. 4. ((?) Linckia dip- 
lax, a single arm broken 
at both ends regenerat- 
ing. (^) Aspect of proxi- 
mal end of arm. 

Fig. 5, Linckia. pacifica, 
regenerating from a sin- 
gle arm, broken off ob- 
liquel}' from the original 
disc; note four new arms 
and disc, the outer arms 
larger than the two inner 

can regenerate a new disc and other arms from an arm torn olf 
without any part of the disc attached, and to the denials by other 
authors that such radical restoration can take place. In the case of 
Linckia diplax there seems to be no doubt of the capacity of an 
arm torn off at some distance from the disc to regenerate a com- 
plete new animal from its proximal surface. The possibility that 
these arm pieces were thrown off by autotomy instead of being 
torn off by enemies may be noted, but such a condition makes the 

Restorative Regeneration of Linckia. 


regenerative phenomena none the less interesting. I have seen no 
example of the regeneration of several new arms (or a new disc 
and arms) from the distal end of a mutilated arm, as observed by 
the Sarasins in Linckia multifera (Ergeb. Naturforsch. auf Ceylon, 
1884-85, I, Wiesbaden, 1888). In all cases of regeneration from 
the distal end of an arm noted among the Apia reef starfishes, 
simply a continuation, in straight line, of the tapering tip oc- 
curred. Among the figures will be noted the illustrations of three 
specimens in which the regenerating arm has had its distal end 

Fig. 6. Linckia diplax, {a) a specimen regener- 
ating parts of two arms: {b) the aspect of a normal 
madreporite (compare with the regenerated madre- 
porite shown in figures i and 2). 

torn off (or thrown off) as well as having been Itself broken off 
from its basal extremity, and thus freed from the rest of the body 
to which it originally belonged. In all of these cases of mere seg- 
ments of a single arm regeneration is proceeding at both mutiliated 

In Figures i and 2 a new mouth and both^ madreporites are in 
the regenerated part. In Figure 3 a new mouth has been already 
regenerated, but no madreporite as yet. In Figure 4 is shown an 

^Linckia diplax is characterized by the possession of two madreporites. 

356 Vei-non L. Kellogg. 

arm torn off at some distance from the disc, just beginning to re- 
generate. The cut end has "calloused" over, apparently by the 
inbending of the edges of the body wall, but in the center is left 
a small opening (serving as mouth ?). No protuberance indi- 
cating new disc or arms has yet appeared. The arm segment, 
which is regenerating at both ends, shown in Figure 5, is of an- 
other species of Linckia, probably pacijica, and had an obliquely 
cut surface at the proximal end, and the two outer arms of the four 
regenerating ones, that is, those nearest the parent arm, are about 
twice as well developed (as far as size goes) as the other two. 
No madreporite is yet developed on the new discal portion. The 
specimen illustrated in Figure i is regenerating but three new 
arms instead of the normally missing four. In all the specimens 
illustrated by Figures i, 2, 3, 4 and 5 the arms were undoubtedly 
broken off without any part of the disc attached. 




In connection with the experimental breeding and rearing under 
controlled conditions of food supply of many lots of silkworms 
{Boinbyx mori) during the last three years, the writers have made 
certain observations and experiments incidental to the main object 
of the investigation, the results of some of which may be here 
briefly abstracted. 

Food Conditions in Relation to Sex Diferentiation. 

It has been assumed by some authors that poor nutrition of de- 
veloping organisms is an extrinsic influence tending to determine 
the sex of the organism to be male and good nutrition an influence 
tending to produce females. The most important part of the as- 
sumption is the idea that sex is subject to control by the environ- 
ment of the organism — that sex is not inherently predetermined in 
the germ. 

From the notes of the writers recording the results of an ex- 
perimental rearing of numerous lots of silkworms on reduced ra- 
tions in 1 901, 1902 and 1903, the following data are extracted 
touching the problem of the relation of nutrition to sex differen- 
tiation. From an inspection of these data it will be noted that a 
test is included of the possible influence of poor nutrition of the 
parents (and grandparents) in determining the sex character (if 
predetermined) of the germ cells, as well as of the possible imme- 
diate influence of nutrition in determining the sex of developing 
individuals. It will be noted also that we have had in mind the 
justly made criticism of most observations on the food and sex 
problem, namely, that no attention is paid in records of an appar- 
ent overproduction of males following poor nutrition, to the deaths 
which ensue before the count is made, and that, as the females (it 
being assumed) actually require more food to complete their de- 

358 Fernon L. Kellogg and R. G. Bell. 

velopment, the preponderance of males is due to the untimely 
death of the females. 

A series of lots of ten individuals each were reared in 1903, 
with the specific intention of testing the assumed influence of nu- 
trition on sex determination. These lots included: (a) a lot un- 
derfed during the whole of larval existence; (b) a lot underfed 
during the second to fifth intermoulting periods, inclusive; (c) a 
lot underfed during the third to fifth intermoulting periods; (d) 
a lot underfed during the fourth and fifth intermoulting periods; 
(e) a lot underfed during the first intermoulting period only; (f) 
a lot underfed during the second intermoulting period only; 
(g) a lot underfed during the third intermoulting period only; (h) 
a lot underfed during the fourth intermoulting period only; (i) 
a lot underfed during the fifth intermoulting period only. From 
the rearing of such lots it was hoped to determine what influence 
reduced rations might have on the determination of sex, and also, 
if any, at what time in the larval life the influence was most potent. 
A consideration of the records of the rearing of these lots at the 
end of the season compels us to say : that the lots were much too 
small to aftord trustworthy generalizations; that dissections of 
the larvae at various ages reveal an unmistakable differentiation in 
sex (indicated by gross differences in the reproductive glands) at a 
time as early at least as the beginning of the third intermoulting 
period, so that experimental lots c, d, g, h and i were distinctly 
superfluous; but finally, that it may be affirmed from the meager 
data afforded by experimental lots a, b, e and /, that the reduction 
of the food supply (this reduction brought as near as possible to 
a living minimum) did not produce any unmistakable results in 
the way of an overproduction of males. 

Data of more interest are those derived from an inspection of 
the records of the experimental rearing of various larger lots of 
silkworms in 1901, 1902 and 1903. In 1901 the records for five 
lots of twenty larvae each may be referred to : 

Lot I — Fed optimum food; no deaths before emergence of 
moths ; produced 8 males, 1 2 females. 

Lot 2 — Fed optimum food; 2 deaths before maturity; produced 
7 males, 1 1 females. 

Notes on Insect Bionomics. 359 

Lot 3 — Fed one-half (approx.) of optimum of food; 4 deaths 
before maturity; produced 10 males , 6 females. 

Lot 4 — Fed living minimum of food; 3 deaths before maturity; 
produced 10 males, 7 females. 

Lot 5 — Fed living minimum of food; 6 deaths; produced 9 
males, 5 females. 

Four lots of twenty larvae each reared in 1902 may be referred 
to.^ These larvae were the offspring of parents of the variously 
fed 1 90 1 lots, and the character of the food supply of the parents 
is indicated as well as that of the larvae themselves. 

Lot I — Fed optimum; born of optimum food parents; no 
deaths before maturity; produced 12 males, 9 females (21 indi- 
viduals in this lot by mistake) . 

Lot 2 — Fed minimum food; born of optimum food parents; 7 
deaths before maturity; produced 8 males, 5 females. 

Lot 3 — Fed optimum food; born of minimum food parents; 1 1 
deaths before maturity; produced 6 males, 3 females. 

Lot 4 — Fed minimum food; born of minimum food parents; 3 
deaths before maturity; produced 11 males; 6 females. 

The records of eight lots of twenty-five larvae each reared in 
1903 may be referred to. The food supply condition of the par- 
ents and grandparents, as well as of the 1903 progeny, are given. 
("O" indicates optimum food, "M" indicates minimum food). 

Lots Fed Parents Grand- before Males Females 

parents maturity produced produced 

I O O O 2 13 10 

2 M O O • 2 14 9 

3 O M O 3 8 14 

4 M M O 6 8 II 

5 O O M o 15 10 

6 M O M o II 14 

7 O M M 20 2 3 

8 M M M 2 1 - 2 2 

^Because of backward season all 1902 larva; were fed for their first 20 days 
(=:rabont one-third of whole larval life) on food of a poor quality, namely, 
lettuce and mulberrv buds. 

360 Venion L. Kellogg and R. G. Bell. 

The writers present these figures, actual data, for what they 
may be worth. Like the data of the smaller lots previously re- 
ferred to, they at least show that individuals living through their 
whole post-embryonic life on the smallest food supply capable of 
sustaining life, a supply varying from M to >^ of the supply nor- 
mally used by individuals of the species, do not necessarily become 
males. Whether the figures indicate an appreciable influence of 
this nutrition on the determination of sex can be determined by the 
readers as well as by the writers. In the rearing season (March 
to June) of this year ( 1904), the writers purpose devoting much 
larger lots of individuals to the continuation of the experiment. 

Forced Pupation. 

Experiments were made to determine how early in larval life 
the food supply could be cut off without stopping the metamor- 
phosis (development) of the silkworm, whether such forced ab- 
breviation of the food-taking period results in any unusual struc- 
tural or physiological modification in the stages which follow the 
withdrawal of food, and whether the metamorphosis (in particu- 
lar, pupation) is hastened when food is withdrawn in late larval 
life, an adaptation often assumed to be possessed by Lepidoptera. 
Such an adaptation would obviously be of real advantage, as it 
might often save individuals from death due to a sudden disap- 
pearance of the food supply, or to a sudden accidental incapacity 
to gain access to the food supply. 

The silkworm spends normally about sixty days in the larval 
(feeding) stage, divided into five actively feeding intermoulting 
periods of about ten days each, by four brief two-day moulting 
periods, during which no food is taken. On the eleventh or 
twelfth day (from 270 to 300 hours) after the fourth moult, the 
larva "spins up" and pupates. 

Twenty healthy silkworms were selected at random from a 
large lot (several hundred) which had been reared in one tray, 
all the individuals, of course, under the same condition of food 
supply, temperature, humidity, light, etc. Of the twenty, one 
was fed as long as it would take food; the other nineteen were de- 

Notes on Insect Bionomics. 361 

prived of food variously from the time of the fourth moult, from 
one day after the fourth moult, from two days after, from three 
days after, and so on until individuals were obtained representing 
a withdrawal of food supply for a period of but a day before the 
normal time of giving up eating to begin spinning, through periods 
of two days before, three days before, four, five and so on to 
twelve days before, the twelve-day period being the whole of the 
feeding period normally lasting from the fourth moulting up to 
spinning time. The following table displays the conditions and 
results of the experiment : 


Vernon L. Kellogg and R. G. Bell. 

O -^^ 

oot- 'bof'boofo'b 


-* 10 

bC be bC be ^ bC be 

1-H (M r^ M jy 10 rf 
(M ^ ^ ^ ^ iO(N 

«4-l 0) 

OOt^ 2cOO0O000l>(M(M g 
C CC CI Pl CI PI fl cj fl-c! 



o 9 »< 

"3 1=1 




»>» ^ ^ ^ C3 ^ ^ j^ 

O 4, 

u to 

T3 O 


O O 


ja o 
'H S 

« a 

p M 

> a 


2 a 


a a a 


03 O) 


2 o 

- - . « ^ ^ . » . (D 

t^ K*^ r*^ ^ ^ r^ r^ „^ 

O fl 


aa a 



G (-'^ 

Q 0^ O; G OJ 
(M(N(N(N(M(N(M(M ----- 

O O 


bo be be 

.a .3 .2 

eeoJcSojojojcSesSaoSy, 9^9^0^000 

%-i o 




(N (N (N IM (N (M rH i-H rH tH ,-1 , 


03 OJ 


a a a a a a a a a a a a a 

a a a a 

1— i^T— 1^1— i^iOiC lOi-ii-H 



Notes on Insect Bionomics. 363 

Note I — Bottom of bottle in which larva was confined was 
lined with threads May 26; threads extended up one side May 
27; threads swung across bottle from side to side May 28, and 
larva actively spinning very damp threads; on May 29 larva was 
spinning a closely woven circular carpet on bottom of bottle, and 
on May 30 larva pupated on this carpet (no cocoon). 

Note 2 — Bottom of bottle lined with threads May 26; larva 
still slowly spinning random threads May 28. 

Note 3 — Bottom of bottle lined with silk coating May 18. 

Note 4 — Larva lined bottom of bottle with stray threads be- 
fore dying. 

Note 5 — Slight progress in the spinning by June 6, P. M.; bot- 
tom of bottle lined with threads. 

From these results it may be said that silkworms may be cut 
off from a food supply nearly seven days before the normal limit 
of their feeding time and yet complete their development (spin, 
pupate and emerge as imago) . These seven days represent a little 
more than half of the last intermoulting actively feeding period, 
or about one-ninth of the whole larval (feeding) life. The depri- 
vation of food for from one to four days seems neither to hasten 
the metamorphosis nor to modify it appreciably, nor to result in 
the production of a moth of lessened size or lessened fertility. The 
larvae deprived of food not more than four days before normal 
close of feeding time do not immediately spin and pupate, but 
wait restlessly for the normal time of pupation (approxim^aMly 
twelve days after the fourth moulting), and then normally spin 
and pupate. If deprived of food for more than four days and 
less than seven, the larvas shorten their last intermoulting stage 
to about seven days, forming, however, a normal cocoon and trans- 
forming into a normal moth. If the larvae are deprived of food 
eight days or more before their normal splnning-up time, they in- 
variably die without forming a cocoon, and in only one case was 
pupation accomplished. A beginning at spinning (see notes) is 
made by larvae fed for more than two days after the fourth moult- 
ing, but no spinning at all is done by larvae deprived of food from 
the day of fourth moulting or from the first or second day there- 

364 Vernon L. Kellogg and R. G. Bell. 

The twentieth larv^a of the lot was to be deprived of food 216 
hours after the fourth moult, but it began spinning up in 200 
hours (eight days) after, and pupated on the following day. Here 
is a normal variation of four days out of the usual twelve of the 
last feeding stage, just about as much shortening as the extreme 
that could be induced by actual deprivation of food. 

Loss of Weight Daring Pupal Life. 

A belief among commercial breeders of silkworms that there 
is a loss in weight of the cocoons (silk) accompanying pupal life 
is indicated by their recognized wish to make an early sale of the 
cocoon product. This loss is generally attributed to "evaporation 
from the cocoon." The question arose as to whether the loss in 
weight of the pupa-containing cocoon might be not a loss in weight 
of silk but an accompaniment of developmental changes in the 
pupa, a process in which stores of nourishment (in the larval body) 
are being converted into moth with chemical changes which might 
occasion some loss in weight. Therefore in four individuals the 
cocoon and pupa were weighed separately once each day from the 
time of pupation to time of emergence of the moth, while at the 
same time the daily weights of the naked chrysalids of three other 
lepidopterous species were determined to see if a loss of weight 
accompanied pupal aging in them as well as in the silkworm 
moth. The following table shows plainly the results of these ob- 
servations : 

Notes on Insect Bionomics, 






(M CO 

q r-; 

^ c*- 


t^ CO 

q r-; 

<M CO 


00 Cs 






>> • 

05 1C 



10 CO 

q i-H 





00 --H 
05 CO 


T— 1 


^ CO 

q r^_ 

q rH 




May 24 
(6 p. m.) 


1— 1 1—1 


-* CO 
q ^ 

r^ CO 

q rt_ 







iM CO 





>> • 

03 &< 

10 iC 
00 (M 




T-l 1— ( 





May 20 
(4 p. m.) 

CO >o 

I— 1 









03 &, 

05 IM 


"* CO 



03' ci 

bi) bJD 



lO CO 

1— 1 




I— 1 














rH 03 

s g 



■< u 

. c3 

-(^ ,— 
_Q CO 






o3 g 



cm e 
^ s 
-2 § 


t-i — ' 



m ^ 


Vernon L. Kellogg and R. G. Bell. 


O 3 

^ CO 


4:J-d : 

^.2 : 


00 CO 

00 .-H 






00 l-H 






to 05 

00 rH 

l-H (M 




O 13 


<£) CO 




June 1 
(3 p. m.) 


00 CO 



05 1-; 


May 31 

(5 p. m.) 

CD (M 


cft CO 

00 .-H 





May 30 
(4 p. m.) 

00 (M 

02 T-H 

<N CO 


^ o 





05 ^• 




<N O 

--H CO 

l-H (M 



















o 5: 


O Sq 


O <o 


03 d 

Notes on Insect Bionomics. 367 

From this table it is apparent that the silken cocoon loses a 
very small amount, about 4 per cent., of its weight in the first day 
after its completion, and then loses no further weight; that the 
pupa loses weight slightly but persistently and steadily from day 
to day throughout its entire duration, the total loss amounting to 
about 14 per cent. ; and that the pupae of three other lepidopterous 
insects, namely, the tent caterpillar {Clisiocampa sp.), checker- 
spot butterfly {Melitaea sp.) , and mourning-cloak butterfly (Eiiva- 
nessa antiopa) also steadily lose weight from day to day, this loss 
being very considerable in two of these species, viz., about 35 per 
cent, in the case of one and 65 per cent, in the case of the other. 




With 2 Plates and 15 Figures in the Text. 

An experimental study of the avian egg has led me to examine 
the following points: 

1. The location of the embryo in the material of the unincubated 

2. The direction of growth before, and after the appearance of 
the primitive streak. 

3. The origin of the material from which the later embryo 

According to Kopsch^ this third point has been definitely 
settled. He concludes that nearly all of the embryo develops 
from the primitive streak. I quote his own words: "Somit 
entsteht der Embryo, mit Aussnahme des praechordalen Teils, 
des Kopfes, durch Umwandlung des Primitivstreifens." 

P have already mentioned some experiments, and will describe 
others, in the following pages, which seem to prove that only the 
trunk and caudal regions of the embryo arise from the material 
of the primitive streak. 

The methods used by Assheton, myself, and Kopsch are practi- 
cally the same, and therefore require little explanation. I have 

^Kopsch, Fr. Ueber die Bedeutung des Primitivstreifens beim Hiihnerembryo. 
Leipzig, 1902. 

^Peebles, F. A Preliminary Note on the Position of the Primitive Streak, and its 
Relation to the Embryo of the Chick. Biolog. Bulletin, Vol. IV, No. 4, 1903. 

370 Florence Peebles. 

again used the method described in my earlier work/ A small 
window was made in the shell just above the blastoderm, and the 
operation performed, after which the opening was closed by a 
piece of shell, sealed with strips of the shell membrane. All in- 
struments used in the experiments were carefully sterilized, and 
the shells of freshly opened eggs used for closing the windows. 
The loss of eggs through infection was small. In most of the 
experiments one egg in each set was opened and then sealed again, 
without operating upon it, in order to have a check with which to 
compare the eggs upon which experiments were made. In this 
way it was possible to determine roughly, after further incubation, 
whether abnormalities were due to opening the egg or to the 
operation performed upon the blastoderm. In general it was 
found that the development of eggs in which windows were made 
was delayed about two to four hours. 


In 1896, Assheton^ described some experiments that he made 
on the unincubated blastoderm of the chick. Sable hairs were 
inserted at various points and their position determined after 
periods of incubation varying from eighteen to forty hours. 
Assheton proved that Duval's^ theory of the formation of the 
primitive streak is incorrect, that instead of forming by the con- 
crescence of the posterior margin of the blastoderm, the primitive 
streak appears in the region of the unincubated blastoderm which 
lies between the center and the posterior margin of the area 
pellucida. I have repeated Assheton's experiments, making the 
injuries with a hot needle instead of a hair, without removing the 
egg from the shell. The results agree with those of Assheton, as 
the following experiments show: 

'Peebles, Florence. Some Experiments on the Primitive Streak of the Chick. 
Archiv. fur Entwickelungsmech. der Organismen. VII Band. 1898. 

^Assheton, R. An Experimental Examination into the Growth of the Blastoderm 
of the Chick. Proceedings of the Royal Soc, Vol. 63, 1896. 

^Duval. De la Formation du Blastoderme dans I'Oeuf d'Oiseau. Annales des 
Sciences Naturelles, Zoologie, Vol. 18. 

The Location of the Chick Embryo. 371 

Experiment I. A small window was made in the shell of an egg 
a few hours after it was laid. The blastoderm measured 2.8 mm. 
in diameter and the area opaca and area pellucida were faintly 
defined. A hot needle (No. 12) was inserted in the center of the 
blastoderm (Text-fig. i, x) and quickly withdrawn. The shell 
was sealed and the egg put in the incubator, the temperature of 
which varied from 37°-39° Centigrade. At the end of twenty 
hours the egg was opened, and the blastoderm killed, removed 
and stained. The primitive streak was clearly defined (PI. I, 
Fig. i) extending from the posterior margin of the area pellucida 
to the point of injury (x). The cells around the wound seemed 
greatly increased in number and showed evidence of forward 
growth which must have been stopped by the injury. 

Other eggs, injured in the same way (Text-fig. i) were left in 



the incubator for a longer period, from thirty to forty-eight hours. 
PI. I, Fig. 2 is a surface view of one of these embryos after forty- 
eight hours' incubation. The embryo is well developed, fourteen 
pairs of somites are present, and the heart is forming. The injured 
area lies dorsal to the heart on a level with the anterior somites. 
The brain region has failed to develop. 

From Assheton's results, and from these just described, we 
must conclude that the primitive streak and the greater part of the 
later embryo form from that region of the unincubated blasto- 
derm which lies behind the center, between it and the posterior 
margin of the area pellucida. The question arises whether or 
not the posterior margin of the area pellucida is a fixed region in 
all eggs, and what the relation of the long axis of the embryo is to 
the long axis of the shell. In Text-fig. 2, an egg is represented 

372 Florence Peebles. 

as opened above the blastoderm. The air chamber, which Hes in 
the blunt end of the shell, is at the left, and the pointed end at the 
right. The chalazae extend on each side of the yolk in the long 
axis of the shell. In this position the blastoderm may be divided 
into right and left halves, the arrow a-b indicating the median 
plane of the bi-laterally symmetrical embryo. We then speak of 
the region c as the anterior border, and d as the posterior border 
of the blastoderm. 

Assheton, in another series of experiments, has made two 
injuries in the unincubated blastoderm (Text-fig. 3) one in the 
center (x) and the other in the posterior border [y) of the area 
pellucida. He found that the primitive streak appeared later 
between these two injuries, and he concluded from this that the 
point (jy) marks the posterior end of the embryo. 

In order to distinguish the region of the primitive streak from 
the rest of the area pellucida, I shall call it the radius x-y. This 
radius with the corresponding one anterior to the center make the 
diameter which represents the median longitudinal axis of the 
embryo. In order to determine the constancy of the occurrence 
of the embryo in this position I have kept the record of 100 eggs. 
The eggs were taken from the nest on the same day that they were 
laid. They were placed in the same position in a basket from 
which they were transferred to the incubator. After incubation 
for eighteen to forty-eight hours the embryo in every fertile egg, 
with the exception of two, was found in the median line (Text-fig. 
2 a-b). The two exceptions are shown in Text-figs. 4 and 5. 
The first embryo (Text-fig. 4) was incubated eighteen hours. 
At the end of this time the primitive streak had formed, but in- 
stead of lying on the radius x-y it extended from the center to the 
right side of the blastoderm and was bent towards the posterior 
margin. The second egg (Text-fig. 5) was incubated for a period 
of twenty-eight hours. The normal embryo lay at right angles to 
the line a-b. In both of these eggs the chalazae were found in 
abnormal positions, and the yolk membrane was wrinkled in many 
places, showing that the yolk had been abnormally twisted in its 
passage through the oviduct. 

After I had discovered that the position of the normal embryo, 
when undisturbed by twisting or shaking, is constant, I deter- 

The Location of the Chick Embryo. 


mined to find out, if possible, whether the embryo would form on 
any other part of the blastoderm if development on the radius x-y 
was prevented. 

Experiment II. The blunt end of the egg was held in the left 
hand so that the blastoderm lay on top of the yolk; a small window 
was made immediately above it, and a series of injuries were made 
with a hot needle in the radius x-y. The number of injuries was 
dependent upon the size of the area pellucida. Usually there is 
space enough to insert the tip of a No. 12 cambric needle in three 
places between x and y (Text-fig. 6) before all the cells are 

At the end of eighteen hours the eggs were killed, but no trace 
of primitive streak in another region was found. About 60 per 
cent, of the blastoderms showed a large hole where the area pellu- 
cida had stretched apart in the growth of the blastoderm. No 

5 6 

evidence of the formation of the embryo around the margin of the 
hole could be found. 

The experiment was repeated and the eggs were incubated 
from thirty to forty hours. An examination of these embryos 
showed no development around the margin of the wounded area, 
but in front of it and posterior to it some development had taken 
place. In PI. I, Fig. 3, a surface view of an embryo of forty hours' 
incubation is given. The brain is abnormal, but shows no lack 
of material, the notochord is present, but greatly reduced in 
length. There is some trace of the heart lying on each side of the 
notochord, but back of it none of the embryo has formed. There 
is no evidence of growth of the area pellucida in a posterior direc- 
tion, but anteriorly it is the normal size and shape. 

Another embryo incubated thirty hours is shown in PI. I, Fig. 
4. In this embryo no brain developed but growth from the heart 

374 Florence Peebles. 

region caudad is evident. The injured area (w) is surrounded 
by thickened ridges and back of the hole made by the wound the 
notochord is present. Behind the notochord Hes the posterior end 
of the primitive streak. No mesoblastic somites are present. 

The object of these experiments vv^as to prevent development 
along the radius x-y by killing the cells in the region w^here the 
primitive streak develops. In this w^ay it was hoped that the 
primitive streak might be formed in some other part of the area 
pellucida. The results show very clearly that no other part of the 
blastoderm is capable of forming the primitive streak. They also 
show that the region of the unincubated blastoderm along the 
radius x-y is the region from which the mesoblastic somites 
develop, /. ^., the trunk region of the embryo. From these ex- 
periments it seems evident that the position of the embryo upon 
the blastoderm is determined before the egg has been incubated, 
and probably before segmentation is completed, for some of the 
eggs which I used were operated upon within two hours from the 
time that they were laid. 


MarshalF describes the growth of the blastoderm from the be- 
ginning of incubation as follows: "After incubation has com- 
menced, the blastoderm spreads rapidly, retaining its circular 
shape. By the end of the first day of incubation it is about the 
size of a sixpence, and by the end of the second day it has 
extended nearly halfway round the egg." 

According to DuvaP the edge of the blastoderm advances over 
the egg at every point except at the posterior margin, and the 
edges on each side of this point meet each other in the middle line 
to form the primitive streak ("plaque axiale"). Assheton's^ 
experiments have proved, however, that the growth is symmetrical 
as Marshall states. 

'Marshall. Vertebrate Embryology. 
^Duval. Loc. cit. 
^Assheton. Loc. cit. 

The Location of the Chick Embryo. 375 

While the margin of the area opaca is symmetrical, that of the 
area pellucida is not. During the first few hours of incubation 
the two areas increase uniformly, but towards the fifteenth hour 
the area pellucida begins to extend posteriorly, the anterior region 
remaining spherical in outline. 

Experiment /. The uniformity of growth in the anterior half 
of the blastoderm can be seen in the following experiment. The 
unincubated blastoderm was injured at three points (Text-fig. 7, 
X, p and 0) ; the needle was inserted at the center (x) at the middle 
point of the anterior margin (/>) and at the right margin of the 
area pellucida (0). The injuries in the margin were at equal 
distances from the center. After eighteen hours' incubation the 
distance between x and p was the same as the distance between x 
and (PI. I, Fig. 5) showing that the lateral and anterior growth 
were the same. The primitive streak was formed, but its posterior 
end (jy) was much further from x than x was from p, while the 
distances before incubation were equal. 

The results of earlier experiments^ led me to believe that the 
region immediately in front of the primitive streak represents an 
area of rapid growth, because an injury made in this region did 
not affect one structure alone, but disturbed the organ covering a 
large area. This is also true when the center of the unincubated 
blastoderm is killed (PI. I, Fig. 2). 

Experiment II. In order to determine the extent of growth in 
an interior direction from the center of the blastoderm I injured 
a point in the center (Text-fig. 8, x) and one on the same level at 
the side (o). The eggs were incubated thirty-six to forty hours. 
PI. I, Fig. 6 is a surface view of an embryo at the end of thirty 
hours. The injury in the center of the blastoderm produced 
great disturbance in the development of the embryo anterior to 
the heart. No forward growth took place in the median line. 
The wound (0) at the side, which did not move forward, is at a 
level with the anterior somites. The normal growth of the margin 
of the area pellucida on the side of the injury did not take place. 
The margin is irregular and a peculiar rod of cells extends from 
the marginal wound to the median line of the embryo. 

^Peebles. Loc. cit. 


Florence Peebles. 

Experiment III. In another set of experiments the two injuries 
were made about .5 mm. further forward (Text-fig. 9, x and 0). 
The position of the injuries after forty hours is seen in PI. I, Fig. 
7. The wound at the side (0) has advanced with the growth of the 
blastoderm but the wound (x) in front of the somites has prevented 
the formation of the head, and the embryo is reduced in length 
anteriorly, the trunk and caudal regions are about the normal 

From these experiments it seems evident that the region in front 
of the middle point of the area pellucida is the seat of active 
growth in an anterior direction. 

Experiment IV. In order to determine the extent of growth 
posteriorly, two injuries were made, one in the center of the unin- 
cubated blastoderm and the other in its posterior margin (Text- 
fig. 3, X and y). The embryos were incubated thirty-six hours to 

two days. They developed somites and medullary folds in the 
area between the wounds. PI. II, Fig. 8, represents a surface view 
at the end of thirty-six hours. Notochord and somites have 
developed between the two wounds. The actual distance from 
X to y before the experiment was i mm. After incubation it was 
3 mm. showing an increase in length of only 2 mm. The normal 
embryos at this age measure 4 mm. from heart to caudal end. 

The results of these three sets of experiments show that the 
embryo may be greatly reduced in length by preventing growth 
anteriorly with the wound x and posteriorly with the wound y^ 
and that the area pellucida grows less rapidly at the sides than in 
the median line. 

Up to this time the experiments which I have described have 
been made upon the unincubated blastoderm. The change in 
the size and the shape of the area pellucida is comparatively 

The Location of the Chick Embryo. 377 

slight before the appearance of the primitive streak when the area 
becomes pear-shaped. 

Kopsch^ has found that when two wounds are made in an em- 
bryo of twenty-four hours' incubation, at a distance of 2 mm., one 
at the anterior, and the other at the posterior end of the primitive 
streak, the embryo does not reach its normal size in later develop- 
ment. The entire body is much shortened, and lies between the 
two wounds. I have repeated this experiment, and have obtained 
the same result, the primitive streak, in the eggs upon which I 
have worked, is much longer (3 to 3.5 mm.) in a twenty-four hour 
chick, and the anterior end is no longer visible, the head process 
and the notochord are present. 

Another series of experiments was made by Kopsch when the 
primitive streak measured about 4 mm. A series of five injuries, 
at 1.5 mm. spaces, were made along the side of the primitive 
streak, and parallel with it. The embryos were incubated fifty 
and one-half hours, and at the end of this time the regions of the 
five wounds were located. Growth in length was greater in the 
region back of the anterior end of the primitive streak than it was 
in front of it. 

I have already described experiments which I have made upon 
the primitive streak and have tried to show that the anterior end 
of the primitive streak of sixteen to eighteen hours represents the 
region of the later embryo which lies back of the heart between the 
anterior somites. 

Experiment V. These experiments were repeated with some 
modifications. Instead of injuring the anterior end alone, a 
second wound was made at the posterior end (Text-fig. 10, x and 
y). The embryo at the time of the operation was from sixteen to 
eighteen hours old. After forty hours a normal embryo developed 
but instead of extending posteriorly to the usual length it was 
shortened 2 mm. Another egg injured in the same way (Text- 
fig. 10) developed into an interesting embryo (PI. II, Fig. 10). 
The posterior wound {y) healed so that no trace of it could be 
discovered, but the anterior wound (x), through the further 

^Kopsch. Loc. cit. 


Florence Peebles. 

growth of the embryo, was left on one side. The only disturbance 
evident was in the medullary folds and somites on the side of the 

Experiment VI. A wound at the posterior end of the primitive 
streak is alone sufficient to shorten the embryo caudad. In PI. II, 
Fig. II, a surface view of an embryo of forty hours' incubation is 
shown. The wound (y) was made at the posterior end of the 
primitive streak of eighteen hours. Fourteen pairs of somites are 
present, and the embryo measures 3 mm. from the heart to the 
anterior border of the brain. This region is normal, but growth 
in a posterior direction has been stopped, by the wound, and the 
length is reduced 1.5 mm. 

Summary. The results from these experiments show that in 
the formation of the third-day chick neither head nor tail region 
can be taken as fixed points, indeed no one point on the blastoderm 
can be said to be fixed. In the series of diagrams (Text-fig. 11, 
A-G) I have indicated, in a schematic way, the method of growth 
from the beginning of incubation until the third day. The growth 
of the area opaca is symmetrical therefore it is not included in the 
diagram. The line I-J represents the plane dividing the unin- 
cubated blastoderm into anterior and posterior halves, and passes 
through the region in the older embryos which corresponds to the 
middle point of the area pellucida before incubation. From this 
point growth proceeds in all directions in the plane of the blasto- 
derm. The growth from the first to the twelfth hour is sym- 

The Location of the Chick Embryo. 379 

metrical. From the twelfth to the eighteenth hour the area pellu- 
cida increases in length posteriorly. From the eighteenth to the 
twenty-fourth hour growth continues posteriorly and also proceeds 
in an anterior direction. From the end of the first day to the end 
of the second day it advances from the heart in both directions, 
more rapidly caudad than cephalad. After this time the tail and 
head are folded off from the surface of the blastoderm. 



I have already spoken of Kopsch's conclusions as to the material 
from which the embryo arises, so that I shall merely mention my 
own results. If, as Kopsch says, the primitive streak represents 
the entire embryo with the exception of the pre-chordal head 
region, then the destruction of definite areas of the primitive 
streak should result in a failure to develop the parts which arise 
from the injured area. 

Experiment I. The first experiment consisted in destroying all 
of the primitive streak except its anterior end (Text-fig. 12). 
This operation is very likely to kill the entire embryo as injury 
to so large an area usually results in a spreading apart of the mar- 
gins of the wound. The further development of an embryo 
injured in this way may be seen in PI. II, Fig. 12. The embryo 
is abnormal, but shows structures which indicate that when de- 
prived of all of its material except the anterior end the primitive 
streak gives rise to the first few pairs of somites; and that the brain 
and notochord develop. The somites are much thinner than in 
the normal embryo. 

Experiment II. In another series of experiments the posterior 
third of the primitive streak was destroyed (Text-fig. 13). 
The destruction of this region resulted in an embryo (PI. II, Fig. 
13), in which the entire caudal region was abnormal. The heart 
and brain, which are not represented in the figure, were normal, 
and fifteen to eighteen pairs of somites were formed in the anterior 
trunk region. This result agrees with Kopsch's view that the 
posterior third of the primitive streak represents the caudal region 
of the embryo from the twentieth somite to the posterior end. 


Florence Peebles. 

Experiment III. In a third series of experiments the middle 
part of the primitive streak was killed (Text-fig. 14), leaving some 
of the material in front, and some back of the v^ound. According 
to Kopsch, in the later embryo the region from the first to the 
twentieth somites should be lacking. 

Nearly all of the embryos which I operated upon, in this way, 
were so greatly disturbed by the wound that all development was 
checked. In PI. II, Fig. 14, a surface view of the body region of 
one of the embryos which developed further is shown. The brain 
and heart were normal, therefore they are not included in the 
figure. Posteriorly the wound {w) stretched apart, but anteriorly 
medullary folds and ten or twelve pairs of somites are present. 
This result indicates that at least ten or twelve pairs of the first 
twenty somites come from the material in the anterior third of 
the primitive streak. 

IS 13 14 15 

Experiment IV. Finally, the anterior third of the primitive 
streak was killed (Text-fig. 15). After further incubation the 
embryo developed a normal brain and heart in front of the wound. 
The trunk region (without the brain and heart) of one of these 
embryos is shown in PI. II, Fig. 15. Back of the wound (w) 
eleven to fourteen pairs of somites are present. By comparison 
with normal embryos of the same age I conclude that these somites 
represent approximately, the tenth to the twentieth pairs, therefore 
all of the somites between the first and tenth pairs have been 
destroyed by injuring the anterior one-third of the primitive 
streak. The notochord is also lacking in these embryos. 

It is evident from these results that the primitive streak of eigh- 
teen hours represents the material from which the trunk and tail 
regions of the later embryo develop; that the posterior third of 
the primitive streak represents the region back of the eighteenth 

The Location of the Chick Embryo. 381 

pair of somites, the middle third represents roughly, from the 
twelfth pair to the eighteenth, while the anterior third supplies 
material for those structures which lie between the heart and the 
twelfth pair of somites, but does not include the chordal region of 
the brain. 


I. The central point of the unincubated blastoderm represents 
the anterior end of the primitive streak, and later, the region just 
back of the heart; therefore, the greater part of the embryo 
develops in the posterior half of the blastoderm. 

2. The region midway between the center of the unincubated 
blastoderm and its anterior border represents the head region of 
the later embryo. 

3. The position of the embryo on the area pellucida is fixed. 
The long axis of the future embryo divides the unincubated 
blastoderm into right and left halves and a line drawn through the 
blastoderm in the long axis of the shell divides it into anterior and 
posterior halves. 

4. Destruction of the material of the unincubated blastoderm 
between the center and its posterior margin does not result in the 
formation of the primitive streak on any other radius. 

5. The growth of the blastoderm is uniform up to the eighth 
to tenth hour, and this uniformity is preserved in the later growth 
of the area opaca, but from the tenth hour the area pellucida 
begins to grow more rapidly in a posterior direction, then later it 
advances anteriorly until it assumes an oval form. Up to the third 
day the region immediately back of the heart (the anterior end 
of the early primitive streak) is the center of growth in all four 
directions, anteriorly, to the left, and to the right, and to a much 
greater extent posteriorly. 

6. Injury to the center and posterior margin of the unincubated 
blastoderm results in a shortened embryo. 

7. Injury at the posterior margin alone will shorten the embryo 
by preventing growth in a posterior direction. 

382 Florence Peebles. 

8. Neither head nor tail region of the embryo can be taken as 
fixed points, the growth at each end proceeds until the head and 
tail become folded off from the blastoderm. 

9. After destruction of all of the material of the primitive streak 
except its anterior end a small embryo with eight to ten pairs of 
somites develops. 

10. The posterior third of the primitive streak furnishes the 
material for the caudal region of the later embryo. The middle 
third represents the trunk region, and the anterior third that part 
of the embryo which lies between the heart and the tenth to 
twelfth pairs of somites. The material of the primitive streak 
does not enter into the formation of the brain. 

The Woman's College, 

Baltimore, June 1, 1904. 

The Location of the Chick Embryo. 383 


Plate I. 

Fig. I. Blastoderm 20 hrs. old. ;f, point of insertion of hot needle before incubation. 

Fig. 2. Ventral view of embryo 48 hrs. old. Injury made in center of unincubated blastoderm 
lies back of heart. Brain undeveloped. 

Fig. 3. Surface view of forty-hour embryo in which the material along the radius x—y had been killed. 
Heart, brain and notochord are present. 

Fig. 4. Embryo 30 hrs. after operation described for Fig. 3. Heart and posterior body region 

Fig. 5. Primitive streak 18 hrs. old. The three black areas indicate the positions of the injuries 
made upon the unincubated blastoderm. 

Fig. 6. Surface view of embryo 36 hrs. old. The black areas indicate the wounds x and made 
in the blastoderm before incubation. 

Fig. 7. Embryo 40 hrs. old. The openings x and indicate the wounds made in the blastoderm. 

Plate II. 

Fig. 8. Embryo 36 hrs. old. The position of the wounds is indicated by the black areas x and y. 

Fig. 9. Embryo 40 hours after injuries were made in the anterior and posterior end of the 18 
hr. primitive streak. 

Fig. 10. Embryo 12 hours older than that in Fig. 9 after the same operation. 

Fig. II. Forty-hour embryo in which an injury had been made in the posterior end of the primitive 
streak of 18 hrs. The black region (y) indicates the wound. 

Fig. 12. Embryo incubated 36 hrs. after four-fifths of the material of the primitive streak was 
destroyed, leaving only the anterior end. 

Fig. 13. Fifty-hour embryo in which the posterior third of the primitive streak (w) was 
destroyed. The brain which was normal is not shown. 

Fig. 14. Surface view of embryo 30 hours after the middle part of the primitive streak was 
destroyed. The normal brain is not shown. 

Fig. 15. Embryo of same age as preceding one. The anterior third of the primitive streak 
was destroyed. Heart and brain which are not given are normal, w in these figures indicates region 
of injury. 



The Journal of Experimental Zoologij. Vol. I 



The Journal of Experimental Zoology. Vol. I 





With 20 Figures. 

The regeneration of a heteromorphic head from the posterior 
end of short cross-pieces of Planaria maculata. Figs. 1-5, led me 
to ex2im'ine Planarta stTnplifissima^ in order to see if the same result 
could be obtained here when short cross-pieces of the worm were 
made. The regularity with which a heteromorphic head can be 
obtained in the latter species when the old head is cut off just 
behind the eyes, Fig. 10, led me to expect that short cross-pieces 
from the body would behave in the same way as do similar pieces 
o( Planaria maculata. The results have proven, however, in part 
otherwise, for while heteromorphic heads do appear on short 
cross-pieces from the anterior regions of the worm, Fig. 11, none 
such develop from the posterior end of short cross-pieces from the 
more posterior regions of Planaria snnplictssitna. On the contrary 
these pieces regenerate a structure from the anterior cut surface 
that appears to be a heteromorphic tail, and another tail from the 
posterior cut surface, Figs. 12, 13. The result is a two-tailed 
and not a two-headed piece. In order to determine if the 
new anterior structure is really a tail, and not simply an undevel- 
oped head, a number of experiments were carried out during the 
winter and spring of 1903-04. 

Before describing the results certain general considerations 
must be spoken of that are intimately connected with the question 

'The principal facts recorded in this paper were reported at the Christmas meeting 
of the American Zoological Society, 1903. 

^This is the same species which, in my earlier papers, has figured as Planaria 
Itigubris. Stevens has recently determined that this worm is P. sim'plicissima. 


T. H. Morgan. 

of heteromorphosis in regions posterior to the old pharynx. An 
important side light is thrown on the problem of axial polarity 
and heteromorphosis by these relations. 

Cross-pieces of Planaria si7nplicissima, that are not too short, 
from the region between the head and the pharynx-chamber regen- 
erate a head on the anterior and a tail on the posterior cut surface, 

Fig. 14. The new pharynx is always situated at the posterior 
edge of the old material. Similar cross-pieces from the region of 
the pharynx-chamber also produce a head at the anterior end and a 
tail at the posterior end. The new pharynx develops in the middle 
of the piece in connection with the old chamber. Cross-pieces 
from the region behind the old pharynx also regenerate a head at 

Regeneration of Heteromorphic Tails. 387 

the anterior and a tail at the posterior end. The new pharynx 
always lies at the anterior end of such pieces, i. e., at the edge of the 
old tissue, and therefore, as it were, in the posterior part of the 
new head that has developed. Fig. 15. It is this relation of the 
new pharynx to the old part that first demands especial considera- 
tion, for, at first sight it is not clear why the pharynx in these pos- 
terior cross-pieces should shift to the anterior end, and not lie, as 
in the more anterior pieces, at the posterior edge. If it did so it is 
obvious that it would appear in a region posterior to that in which 
the normal pharynx lies in the old worm, and it seems that this 
cannot take place. The posterior cut surface can form only that 
part of the tail that lies behind it in the old worm. The anterior 
cut surface can also produce all that lies in front of it in the old 
worm, including the pharynx, although the proportionate distances 
apart of the new structures may be at first very different from 
those in the adult or embryonic worm. We touch here on one of 
the fundamental questions of polarity to which I shall hope to 
return at another time. 

If a heteromorphic tail were to develop on the anterior cut 
surface of a short posterior cross-piece, what should we anticipate 
in regard to its relation to a pharynx .? Should we expect to find 
a pharynx in the new tail turned in the opposite direction, /. e., 
pointing towards the tip of the tail .? But if the new structure at 
the anterior end is a heteromorphic tail why should it develop 
a pharynx at all, since this never develops at the posterior 
end of cross-pieces from this region .? Should we not rather 
expect a heteromorphic tail to behave in this respect in the same 
way as the orthomorphic structure .^ This appears to me to be 
the correct point of view and the results of experiment seem to 
bear out this anticipation. 

Let us apply the same point of view to the regeneration of a 
pharynx in heteromorphic heads from cross-cut pieces of the ante- 
rior regions of the worm. It has been pointed out above that the 
new pharynx appears at the posterior end in case a tail develops 
at this end. Suppose, however, a heteromorphic head instead of 
an orthomorphic tail develops at the posterior end. It is clear, 
from our point of view, that no pharynx should develop, and I 


T. H. Morgan. 

have found that none such is present as a rule. In the few cases 
in which a pharynx appeared in the middle of the piece, the piece 
may have come from the region near to or through the pharynx- 
chamber of the old worm. 

Turning now to the results of regeneration of very short cross- 
pieces of Planaria stmplicissima, it was found that double headed 
pieces are sometimes obtained from the more anterior regions, 
Fig. II, as in P. ?naculata, Figs. 2-5. When short cross-pieces 

of P. simplicissima are cut off posterior to the pharynx-chamber a 
number of them produce a head at the anterior end and a tail at 
the posterior end, especially if they are rather long. Fig. 15; but 
short pieces and sometimes some of the longer pieces also produce 
quite often a pointed structure at the anterior end. In the ma- 
jority of cases these anterior structures never develop into any- 
thing different, and resemble a tail in all respects. In a few cases, 
however, the pointed structure may become a head after some 

Regeneration of Heteromorphic Tails. 389 

time. The possibility that all the anterior pointed structures 
may be only undeveloped heads must therefore be given seri- 
ous consideration. That they are not such in many cases is 
shown, I think, by the following facts. In the first place the 
movement of a piece in which the anterior head is undeveloped 
is very different from that of the two-tailed pieces, and reveals the 
nature of the new part; for, while the former crawls forward as do 
the pieces when first cut from the worm and as do those that 
develop an anterior head, the two-tailed pieces remain fixed in one 
place in the dish, and, if disturbed, fail to move in any definite 
direction. This is what we should anticipate if two tails were 
present working in opposite directions. 

In the second place an orthomorphic pharynx appears as a 
rule when the head is delayed in its development while none such 
appears in the two-tailed pieces. In the third place the peculiar 
motion of the anterior end when it is irritated is similar to that of 
a tail and not like that of a head. Finally, the development in 
one case to be mentioned below of a two-tailed piece with pharynx 
in each tail shows, beyond a doubt, the possibility of the develop- 
ment of a heteromorphic tail in these worms. 

After the short cross-pieces have been cut off for some time it 
is difficult to distinguish the anterior from the posterior end and to 
know which is the anterior heteromorphic and which is the normal 
posterior orthomorphic tail. In order to distinguish these apart, 
I cut off pieces obliquely at one end; in one set the anterior end 
being the oblique one. Fig. 16, in the others the posterior, Fig. 18. 
This necessitated increasing somewhat the length of the pieces and 
brought about in consequence an increase in the number of the 
pieces that regenerated a head at the anterior end. The record of 
one set of experiments of this sort is given here. 

On May 7, twelve short cross-pieces were cut off just behind the 
pharynx. The anterior end of each was oblique. On May 20 
there were alive three two-tailed pieces; the rest having died. 

Twelve short tail-ends regenerated new tissue at the anterior 
end which in two cases at least appeared to be tails. 

In another series twelve short cross-pieces were cut off behind 
the pharynx. The anterior end was square and the posterior end 

390 T. H. Morgan. 

oblique. Of the eight survivors all had pointed anterior and 
posterior ends. 

The pieces just behind the last set (with oblique anterior ends) 
gave five two-tailed pieces, and one piece with head a.nd tail. 

The tail-ends of this set had the posterior tip cut off. The three 
that remained alive developed a pointed anterior end. 

In another series like the last, the first pieces produced six two- 
tailed forms; the second six two-tailed forms; and the tail-ends five 
two-tailed forms. 

In several other cases I allowed one end of the piece to close 
for two or three days before cutting it off. In this way the mor- 
tality of very short pieces, which is otherwise very great owing to 
their immediate disintegration, is lessened. Some of these pieces 
also gave two-tailed forms. 

In only one case did I obtain a piece in which a pharynx was 
present in each tail, and in each turned outward toward the tip 
of the tail, as shown in Fig. 12. The exact location of this piece is, 
I am sorry to say, uncertain. It came, in all probability, from 
the region just behind, or including some of, the old pharynx 
region. I am inclined to think that the latter is the more probable 
location, since the cut may sometimes include somewhat more or 
less than is intended. The direction taken by the pharynges in 
these pieces shows beyond a doubt that one of the two tails is a 
heteromorphic structure, and this lends support to the interpre- 
tation that I have given to the other cases, in which there is no 
pharynx in the heteromorphic tail, and where none should be 
expected to be present on theoretical grounds. 

Two other kinds of apparently heteromorphic structures have 
been met with in carrying out these experiments. In one case a 
piece of P. maculata, whose posterior end was cut off very 
obliquely, regenerated one head on the anterior cut surface and 
another also on the right side of the posterior cut surface, as shown 
in Fig. 8. A tail also developed on the posterior cut surface at the 
left side, which is also the more posterior end of this surface. In 
this case we must look upon the long edge of new tissue on the 
posterior surface as producing a head at one end and a tail at the 
other, very much as occurs when a longitudinal piece of the worm 


Regeneration of Heteromorphic Tails. 391 

is removed. The case recalls those in which the worm is split 
from the posterior end far forward and a head develops at the 
anterior end of the cut surface on one or on both sides. I have 
discussed the meaning of this case elsewhere.^ In a strict use of 
the word heteromorphosis, as I have tried to use it for purposes 
of greater clearness, neither this case nor that of the split-worm 
can be looked upon as an example of axial heteromorphosis, since 
the result depends largely, apparently, on the new part alone 
without relation to the old, and the head and tail are orthomor- 
phic from this point of view. 

In some other cases in which the anterior end is very oblique, 
two structures appear on the anterior edge, as shown in Figs. 19- 
20. One of these is a head and the other appears from its struc- 
ture and movements to be a tail. If so, the case is comparable to 
the last one, and shows the converse condition. Here the tail on 
the side of the anterior cut surface cannot be looked upon as an 
example of axial heteromorphosis. It is rather an orthomorphic 
structure, since it stands in this relation to the remainder of the 
new material on the anterior cut surface. Both cases, however, 
present something of a paradoxical relation. 

The results described in the first part of this account recall cer- 
tain conditions that I have recently described in connection with 
the regeneration of Dendro caelum lacteum. It had been shown 
by Lillie that posterior pieces cut off just in front of, or through, 
or behind the pharynx-chamber do not regenerate an anterior 
end. A histological examination of the anterior end of such 
pieces showed me that a certain amount of new tissue is formed 
at the anterior cut surface, and it was not apparent why the re- 
generation should not go further and produce a new anterior end. 
The results with Planaria simplicissima suggest, although they by 
no means prove, that the anterior part that regenerates in Dendro- 
ccelum may be a heteromorphic tail. For the present, however, 
I wish to leave this question open, until further work reveals the 
nature of the anterior part in this worm. There are some gen- 
eral considerations in connection with the problem of polarity 

^Regeneration. 1900. 

392 1 . H. Morgan. 

and of heteromorphosis that may be very briefly touched upon at 
this time. Although I have not hesitated in earlier papers to 
speak of polarity as a factor in regeneration, I have always 
tried to be careful to state that we are really entirely ignorant in 
regard to its nature. When we see the polarity suddenly reversed 
in cases of axial heteromorphosis it appears that this ought to 
throw some light upon the nature of the factor itself, yet despite 
the numerous surmises that have been made of a material, — chem- 
ical, or electrical nature — we still remain totally in the dark as 
to what factors determine the stereometrical relations of the new 
part. The following facts appear, nevertheless, to have an im- 
portant bearing on this topic, and while they do not offset an imme- 
diate solution of the problem, yet they may point in the direction 
in which an analysis may ultimately be undertaken. 

In the more highly specialized forms the question of what re- 
generates appears, in part, to be connected with the nature of the 
material, or with the kinds of the material that give rise to the 
new cells, and the relation of direction is less apparent. The tail 
of a tadpole regenerates only a tail, even at its anterior end. The 
same appears to be true for the leg of the salamander from certain 
results that I have obtained, which are as yet unpublished. In 
the earthworm as shown by Morgan^ and by Dimon^ the 
regeneration of an orthomorphic head is connected with the 
presence of the anterior structures of the worm, while from the 
part containing the intestine — including by far the greater length 
of the worm — only a tail is, as a rule, regenerated, even from the 
anterior cut end. In these cases it appears that the nature of the 
material must decide the character of the new part, and the polar 
relations do not come conspicuously to the front, although that 
something of the sort still enters into the problem is shown by the 
slower rate, and, in some cases, by the less perfect form of the 
heteromorphic growth. 

On the other hand, in less specialized forms the polar relations 
appear to play a more conspicuous role. In Lumbriculus a head 

'Anatomischer Anzeiger. Bd. 15, 1899. 

^Journal of Experimental Zoology. Vol. I, No. 2, 1904. 

Regeneration of Heteromorphic Tails. 393 

may regenerate from an anterior cut surface, and a tail from a 
posterior cut surface throughout a very considerable region of the 
body. In planarians and in hydra similar facts are known. That 
the specification of the tissues or parts plays a role even in these 
cases is probable, as shown by the cases of heteromorphosis that 
I have described. To many writers it has seemed that the factor 
of polarity maybe something in the nature of a crystallizing force — 
to use the nearest analogy at hand — a sort of perfecting or com- 
pleting principle. Newer results have modified our ideas as to 
this form of explanation, if such an analogy can be called at all 
an explanation. The fact, for example, that in the earthworm 
and in planarians the new head may be very short in comparison 
to the part that is missing indicates that a completing force cannot 
be acting from the cut surface forwards, but whatever the nature 
of the factor it must in large part work from without (surface) 
inward (/. e., toward the cut end). This point has been already 
urged by myself, and by Driesch. 

It is very significant, I think, to find that in planarians the 
shortness of the piece is a factor that enters into the problem as 
to the character of the new part. I have suggested tentatively 
that this means that in Planaria maculata the tendency is stronger 
for the new structure to become a head than a tail, and that 
when the influence of polarity is removed a head appears on each 
end of short cross-pieces. In other worms, as in Planaria sim- 
plicissiTua, the tendency in certain posterior regions to produce a 
tail is stronger than that to produce a head, and two tails appear 
when the polarity is reduced or removed. Why should the length 
of the piece be so important a factor .? Can it be that there is 
a greater difference, chemical or physical, between the two ends 
of a longer piece, so that a stronger polarity is present ^ In short 
pieces, from this point of view, the ends being near together are so 
much alike that the polarity is correspondingly reduced, and, 
under these conditions, the specification of the material of the old 
part is not sufficiently strong to determine the nature of the new 
part. These and many other equally obscure questions remain 
for future investigation to explain. 



With 5 Figures. 



I. Description of C. palma. 

II. Habitat; food. 

III. Activities of the polyp. 

a. Muscular movements. 

b. Geotropism; axial ceUs. 

c. Locomotion; amoeboid cells. 

d. Circulation; cilia. 

IV. The young hydroid. 


This is one of a series of papers which deal with some of the 
phenomena of growth, differentiation and development in 
Corymorphay from different points of view. The study of de- 
velopmental mechanics has long since ceased to consist merely in 
an analysis of the development of egg and embryo. That 
regenerative development must be included goes without saying; 
and it is with the feeling that the normal activities not ordinarily 
considered in the category of developmental processes should be 
included also, that I have incorporated much of what may at first 
sight appear to be purely physiological material. 

Corymorpha is an exceptionally attractive basis for such an 
investigation. In the first place, it combines remarkable powers 
of regeneration with a simple development from the egg, the non- 

396 Harry Beal Torrey. 

sexual origin of sexual individuals, and powers of movement that 
are unusual for a hydroid. In the second place, very little is 
known of its biology, and its value for experimental work has not 
yet been generally realized. No account of its regeneration has 
been published, with the exception of a brief reference to it in a 
former paper by the present author ('02, p. 41). Up to this time 
our knowledge of the egg-development has been based upon three 
stages described and figured by Allman ('71):^ the sessile planula, 
the polyp with six proximal tentacles, and the polyp with sixteen 
to twenty. These Allman took for stages in the development of 
what he called "frustules," minute bodies cut off from the pro- 
cesses that develbp near the base of the stem and really give rise 
to the filaments of the hold-fast. Had he seen the eggs on the 
medusae in his aquarium, this pardonable error would not have 
been made. Agassiz ('62) and Allman ('63) have given brief 
accounts of the natural history of the hydroid and the develop- 
ment of the medusa. 

These works, with several of a taxonomic character, and a 
recent paper by May ('03), comprise the scant publications on 
Corymorpha relating to the subject of this paper. 


The nutritive polyps of C. palma are solitary. The stem may 
reach the length of ten centimeters, tapering gradually from a 
diameter of perhaps six millimeters, near the base, to a narrow 
neck which supports the hydranth. It is covered for about its 
proximal third with a thin, non-supporting layer of perisarc. 
Within is a solid axis of immense vacuolated cells. These have 
almost obliterated the cavity of the stem, which persists as a num- 
ber of small, longitudinal canals lying immediately under the thin 
mesogloea, and usually made conspicuous by their green tinted 

'A Monograph of the Gymnoblastic or Tubularian Hydroids. London, 1871. 
The section in this magnificent monograph devoted to Corymorpha is a reprint of 
Allman's paper in the Annals and Magazine of Natural History for January, 1863, 
p. 1, with but slight verbal changes and the addition of figures. 

Biological Studies on Corymorpha. 397 

The hydranth has a single whorl of eighteen to thirty proximal 
tentacles with a spread of more than twenty-five millimeters. 
The proboscis, terminating in the mouth, is crowned with forty to 
sixty distal tentacles. Just within the proximal tentacles are 
several peduncles which bear numerous medusoid gonophores. 

The stem is anchored by a tangle of filaments which arise on 
the longitudinal canals, beneath the perisarc, usually in pairs. ^ 

II. habitat; food. 

Corymorpha palma is a semi-tropical species, dwelling farther 
to the south than any of the other North American species of the 
genus. It has been found as yet only in two localities: in San 
Diego and San Pedro harbors, both on the southern coast of 
California. It lives under similar conditions in both places. At 
San Diego it was found in a slough near the mouth of the harbor, 
on a muddy bottom which was exposed at mean low water. At 
San Pedro it has flourished at various points in the harbor, always, 
however, on muddy flats. It occurs usually in definitely circum- 
scribed patches, which change their position apparently with 
much caprice from year to year. A favorite location is along some 
small stream that drains the mud flats as the tide ebbs. 

Copepods are numerous on the mud, which often carries 
patches of green composed of diatoms and other chlorophyl- 
bearing protista. All of these organisms seem to serve as food for 
Corymorpha, though the copepods form the staple article of diet. 

III. activities of the polyp. 

Corymorpha captures crawling diatoms and copepods by bend- 
ing its column in a half circle and sweeping the sand with its 
tentacles. Floating organisms are caught when the column is in 
its usual erect position, with proximal tentacles fully extended. 
The oral tentacles are almost always active, bending restlessly 
now outward, now inward, now moving simultaneously, now 
independently. The proboscis is extremely mobile, capable of 
lengthening into a narrow stalk or contracting into a sphere, or 

'For a diagnosis of tlie species, see my paper just referred to. 

398 Harry Beal Torrey. 

turning inside out, or carrying the mouth with its prehensile 
tentacles to the bases of the proximal tentacles on all sides. The 
proximal tentacles are comparatively quiet. For minutes at a 
time they may be held in full extension, motionless, curved grace- 
fully back from the proboscis, on a vertical stalk. Now and then 
one may twitch toward the mouth. Occasionally all may wave 
inward together, grasping the proboscis tightly. The points of 
the tentacles may be, but usually are not, carried directly toward 
the mouth. 

a. Muscular Movements. 

So far as I am aware, the reactions of hydroids (with the ex- 
ception of Hydra) to different sorts of stimuli have never been 
studied. Medusae, on the contrary, have been the subjects of 
extended investigations by Romanes ('76, '']']), Eimer ('78), whose 
paper I have not seen, and Nagel ('93, '94)- In certain respects, 
Corymorpha and some of the craspedote medusae (Carmarina, 
Sarsia) respond similarly to similar stimuli. For instance, the 
proboscis of each may move toward a point of stimulation not on 
it; and increasing the stimulation of a tentacle may increase the 
number of tentacles taking part in the response, and leads finally 
to the contraction of the body of the animal. Hydra and 
Corymorpha^ however, resemble each other more closely in their 
responses than either resembles a medusa. In general, similar 
structures respond similarly, but the tentacles of neither Hydra 
nor Corymorpha react to odorous substances, while according to 
Nagel ('93), the tentacles of Carmarina hastata do. Such excep- 
tions, coupled with obvious differences in structure and habits 
between polyp and medusa, make it necessary to treat each case 

The large size of Corymorpha makes it an unusually favorable 
object for experimentation in this direction. Experiments with 
mechanical, chemical and thermal stimuli brought out the follow- 
ing facts : 

Mechanical Stimuli. Each proximal tentacle responds to a 
touch or pinch from forceps by contracting in the same direction 
with the same strength, whether the stimulus be applied at the 

Biological Studies on Corymorpha. 399 

base or the tip, on the oral, aboral or lateral surfaces. The re- 
sponse is always a bend inward, never outward. In this respect 
it differs from the tentacular responses in some anemones (Cri- 
brina, Sagartia), where a tentacle commonly reacts to a slight 
touch by bending sharply at and toward the point stimulated. 
This reflex is clearly advantageous to the anemone, which it 
enables, to a limited extent, actually to pursue its prey. It is 
supplemented by another. As soon as the tentacle, which is 
adhesive, seizes the stimulating object, it contracts, carrying its 
capture to the mouth, over which it bends. Then, by means of 
the cilia with which the tentacle is covered, the object, at least if 
available for food, is swept off the end of the tentacle and dropped 
upon the lips. 

While the general direction of the movement of the tentacles 
does not vary, the intensity of the contraction varies with the 
intensity of the stimulus. A touch or slight pinch produces a 
waving of the tentacle toward the proboscis, though without 
reaching it; and the tip of the tentacle is not directed toward the 
mouth. A stronger stimulus may cause the tip to touch the distal 
tentacles, may even cause the tentacle to coil against the proboscis. 

When the stimulus reaches a greater intensity, it may induce 
simultaneous movements in several or all the proximal tentacles. 
Before this point is reached, however, it is able to set up move- 
ments in the distal tentacles and proboscis. If a proximal ten- 
tacle contracts, an effect is often evident among the distal ten- 
tacles, even though the proximal tentacle has not touched them. 
This effect is manifested either by a simultaneous downward 
movement or an indeterminate waving of all the tentacles, or by a 
downward motion of a few nearest the tentacle stimulated. 

These movements of the distal tentacles may occur without any 
apparent movement in the proboscis. If, however, the stimula- 
tion of the proximal tentacle is increased (occasionally a very 
slight stimulus is sufficient to produce the movement) the pro- 
boscis also may bend, carrying the mouth and distal tentacles 
toward the tentacle stimulated. This is a coordinated reflex of 
the same purposive aspect as the movements of the proboscis of 
the medusae Sarsia, Tiaropsts (Romanes, '77) and Carmarina 

400 Harry Beal Torrey. 

(Nagel, '94) toward stimulated points on the sub-umbrella. 
While the excitation may be transmitted by means of the nerves of 
the tentacles and proboscis, certain facts indicate that the direct 
pull of the tentacle on the base of the proboscis serves at least to 
reinforce the impulse and aid in guiding the tentacles and pro- 
boscis in the proper direction. For example, the proboscis never 
bends until the stimulated tentacle contracts, although this con- 
traction may be delayed half a second or a second after the 
stimulus is applied — an unusual reaction time; it does bend, how- 
ever, immediately upon the contraction of the tentacle. Again, 
when a simple grip of the forceps does not cause a movement of 
the proboscis, the movement may be induced by adding to the 
tactual stimulus a definite tension stimulus by pulling the tentacle 
or preventing it altogether from shortening. 

Not only may stimulation of a proximal tentacle be followed by 
movements of distal tentacles and proboscis, but by movements 
of the stem as well, which contracts strongly when the stimulation 
is vigorous. Only that part of the stem ordinarily contracts 
which is not invested with perisarc, though the latter is too thin to 
be an effectual hindrance to contraction in the basal region, which 
at times does shorten considerably. 

When one distal tentacle is pinched, the response is similar to 
what occurs when the disal tentacles respond reflexly to a stimu- 
lation of a proximal tentacle; that is, several or all the distal ten- 
tacles may wave outward and downward together, or indiscrimi- 
nately outward. In response to this stimulus, the movement is 
always away from, never toward the mouth; in this respect it is 
contrary to the direction of the movement of the proximal ten- 
tacles. After the first reaction, however, the tentacles may move 
actively and singly toward and away from the mouth. This is 
the characteristic reaction when the stimulation is prolonged. 
The presence of a large food organism in the proboscis cavity will 
cause such movements, which will persist until it has been 
entirely swallowed. They are only moderately efficient, for the 
outward movement of each tentacle is quite as strong as the in- 
ward movement, and the tentacle retains its hold on the captured 
organism for an instant only. They are indeed far less efficient 

Biological Studies on Corymorpha. 401 

than the tenacious tentacles of anemones with their more definite 

When the proboscis is pinched at its base, it bends toward the 
point of stimulation, the distal tentacles waving. A pinch at any 
point, of a sufficient intensity, will induce characteristic move- 
ments of the distal and proximal tentacles and a shortening of the 
stem. If the stimulus is not too strong, only the proximal and 
distal tentacles in the vicinity of the point of stimulation will react 

A slight stimulation of the stem may produce characteristic 
movements of both sets of tentacles. The effect is not related to 
the position of the point stimulated; none of the parts reacting 
appear to distinguish the direction from which the impulse comes. 
The stem may also shorten, even to half its original length. The 
shortening usually takes place in the distal naked portion. The 
proximal third, however, which is covered with perisarc, may also 
contract; the perisarc is very delicate and in no way interferes with 
this or any other movement of the stem.^ 

Chemical Stimuli. Several substances were used: flesh in the 
shape of pieces of the shore gastropods Littorina and Acmcea, and 
boiled ham; clove oil, alcohol and acetic acid. In no case did the 
meat juices have the slightest appreciable effect on the hydroid; 
the same may be said of the clove oil. Only when touched by a 
stream of strong alcohol or acetic acid from a pipette did tentacles 
or column respond; the acid killed the former almost instantly. 
This response is evidently of the tactile order — as when an 
irritating fluid is poured on the hand. Substances which have 
for our perception odors and flavors, appear to produce no 

^The proportion of stem covered by perisarc is based on measurements of 
expanded individuals, under normal conditions. When a hydroid has been standing 
in the same water for a week or two, it usually becomes much attenuated, and the 
part of the stem invested with perisarc then often appears longer than the distal 
naked portion. Often the ratio of the covered to the naked portion of the stem may 
become that of the larva (Fig. 5). 

^Loeb ('95) has already criticised the use of the words "olfactory" and "gusta- 
tory" to describe the reactions to chemical stimuli of animals of whose consciousness 
we are as ignorant as we are of the consciousness of the Coelenterata. 

402 Harry Beal Torrey. 

Thermal Stimuli. A rapid rise in temperature of several de- 
grees, caused by flooding the hydroid gently with warm water from 
a pipette, produced a general contraction of the same character as 
the response to a strongtactuai stimulus.^ Gradual changes in tem- 
perature aflPect both irritability and the rate of growth, increase of 
temperature resulting in increased irritability and more rapid 
growth, and vice versa; the limits, however, were not determined. 

The reactions of Corymorpha to the various stimuli considered 
above may be summarized as follows: All parts of the hydroid 
are very sensitive to mechanical stimuli, irritating chemicals and 
abrupt increases of temperature. Proximity to odorous sub- 
stances, especially flesh, which might serve as food, awakens no 
appreciable response until the substances are actually touched. 
Food organisms, therefore, are probably detected only when they 
strike the hydroid. The mechanism for capturing them is inter- 
esting on account of the definite but dissimilar responses of the 
two sorts of tentacles and the coordination exhibited in the activi- 
ties of all the parts. The proximal tentacles with their great 
spread (which sometimes almost equals the length of the stem) 
serve as the chief means of advertising the presence of food and 
carrying it to the mouth. These functions are sufl&ciently well 
discharged b)^ a movement in one direction only — toward the 
mouth; but the absence of the preliminary movement in the direc- 
tion of the stimulus, which has been noted among the anemones, 
entails a certain loss of efficiency. This loss of efl&ciency is com- 
pensated for to some extent by the movements of the distal ten- 
tacles and the proboscis. The stimulus which causes the move- 
ments is in the great majority of cases liable to be applied to the 
proximal tentacles, on account of their relatively much greater 
spread. And apparently because of this, whether acquired by 
habit or selection, the first movements of the distal tentacles in 
response to direct or indirect stimulation are downward and out- 
ward, toward the proximal tentacles; that is, toward the usual 

^This contraction is the result of muscular activity, does not concern the axial 
endoderm (to be especially considered later under Geotropism) and is not to be com- 
pared, therefore, with such growth processes as were shown by True ('95) to follow 
in radicles of seedlings, transference from water at 0°C. to water at 18°-21° C. 

Biological Studies on Corymorpha. 403 

point of stimulation and away from the mouth. These move- 
ments, together with the tendency of the proboscis to bend toward 
the point of stimulation, carrying the distal tentacles with it, 
undoubtedly supplement the movements of the proximal tentacles 
in bringing food to the mouth, and raise the average of efficiency 
of the prehensile mechanism. 

The contractions of the stem muscles, determining a limited 
range of movement for the hydranth, may be of advantage to 
Corymorpha. The rapid shortening of the stem following strong 
stimulation, however, can have no value as a part of the 
mechanism of prehension, nor does it have any apparent useful- 
ness as a means of defense against predatory enemies. 

h. Geotropts?n; Functions of the Axial Endoderm. 

Up to this point we have been considering the effects produced 
by the contractions of muscles, "^ in tentacles, proboscis and column, 
under certain sorts of stimulation. 

We may now consider another type ot motor reaction induced 
by another sort of stimulus which appears ultimately to affect 
another tissue element. This is the tendency of the stem, in 
assuming its most characteristic attitude, to turn directly away 
from the center of the earth, by what seems to be a change in 
the turgidity of the axial endoderm cells incited by the stimulus 
of gravity. 

It will be unnecessary to enter into an extended discussion of 
the phenomena of geotropism," which are familiar to all. A few 
words will suffice for the purposes of this paper. 

^There are both longitudinal (ectodermal) and circular (endodermal) muscle 
fibers in both proximal and distal tentacles, proboscis and column. As might be 
expected from their activities, the circular fibers in the proximal tentacles form a 
much weaker sheet than the longitudinal, except where each tentacle joins the body 
of the hydranth. There the circular fibers are aggregated into a strong bandlike 
sphincter, and there the tentacles are wont to break away from the hydranth under 
unfavorable conditions. Such a habit of casting the tentacles seems to be character- 
istic of certain anemones, notably Bolocera, and is accomplished by a similar 

^Davenport has distinguished between the responses of free and fixed organisms 
to gravity, following Schwartz in applying to the former the term "geotaxis," and 
applying to the latter the term " geotropism." With the facts which follow in mind, 
it will be difficult, I believe, to see any advantage in this distinction, and it has 
accordingly been disregarded. 

404 Harry Beal Torrey. 

Gravity affects both free and fixed organisms, and, so far as we 
are concerned with it, determines orientation, direction of loco- 
motion, and direction of growth. In free organisms, orientation 
may or may not be accompanied by locomotion. Davenport has 
cited the infusorian Spirostomum as an organism which may 
belong in the latter category. In this case, orientation is finally 
due to the action of cilia with which the animal is clothed; if loco- 
motion is associated with orientation here, it is very slight and 
inconspicuous. Cerianthus, whose negative geotropism was first 
considered by Loeb ('91), orients itself by means of muscular 
action. Though a free organism, it pursues a sedentary habit. 
The same tendencies are manifested by sand-dwelling anemones. 

Among fixed forms may be considered (i) those which are 
attached aborally, but are also capable of some degree of loco- 
motion, such as most of the anemones and the hydroid Corymor- 
pha; (2) those which are permanently attached, such as most of 
the hydroids; and with these must be classed plants, especially 
seedlings. I have recently referred ('04) to the geotropism of the 
anemone Sagartia davisi, the orientation being accomplished, as 
in Cerianthus, by muscles. In the discussion of the geotropism 
of Corymorpha to follow, it will be shown that the orientation of 
the column is probably accomplished, not by muscles, but by 
means of growth processes comparable with the growth processes 
responsible for the orientation of seedlings and, presumably, 
of geotropic hydroids. 

That the characteristic position of the stem in Corymorpha is 
not due to a difference in the specific gravities of the distal and 
proximal regions is apparent when it is seen that not only is the 
hydroid both proximally and distally heavier than water, but 
distally it is heavier than it is proximally. If a hydroid is placed 
in a jar of water after having been slipped out of its proximal 
investment of perisarc, weighted down as that is by sand clinging 
to the filaments of the hold-fast, it sinks at once, hydranth first, and 
lies upon the bottom until the proximal end becomes attached. 
When this occurs, the stem begins to rise, and in an hour is erect. 

The result is in the end the same whether the hydranth is 
present or absent, whether the stem is cut so that the proximal 

Biological Studies on Corymorpha, 


portion is two-thirds, one-third, or even one-eighth the original 
length of the stem. Evidently the stem is generally responsive to 
the geotropic stimulus. The only difference lies in the time con- 
sumed in reaching the vertical position. The longer the stem the 
shorter the period. 

The result is also in the end the same whether the hydroid is 
hung vertically upside down, by the proximal extremity, or right 
side up, by the "neck," just below the hydranth, or by the middle. 
And it matters not whether the hydranth and foot are both or 
either one present or absent. 

Fig. 1. 
Three vertical stems cut at different levels, which were parallel with S one hour 

. Numerous experiments justify this summary. Typical cases 
will be described. To exclude the possible influence of light and 
oxygen on the direction of orientation, the hydroids were com- 
pelled to orient themselves in sealed jars quite full of water, which 
were placed in dark closets. Check experiments in the light and 
in open aquaria gave identical results. Further, about a dozen 
individuals were subjected for three hours to light coming from but 
one direction, without any observable result on their orientation. 

4o6 Harry Beal Torrey. 

Several uninjured hydroids and three (Fig. i) which had been 
cut respectively two-thirds, one-third, and one-eighth the length 
of the stem from the proximal end, all weighted with sand as 
usual, were placed in a dark closet in a jar full of water. As soon 
as they became erect, the jar was tilted at an angle of forty-five 
degrees, the stems of the hydroids remaining parallel with its 
sides. In an hour all the stems had become erect. The distal 
pieces cut from the three mutilated stems were still lying on the 
bottom, unable to rise for lack of hold-fasts. The jar was brought 
back to the vertical, and in another hour the stems had swung 
through forty-five degrees to the vertical, in the opposite direction. 
The time required for such changes varies considerably, but in the 
same experiment the shorter the piece the longer the time — ten to 
twenty minutes longer in the case described. The movement was 
constantly toward, never away from, the vertical position finally 
assumed, and did not suggest in the slightest the method of trial 
and error (Jennings, '04). 

The time required for an inverted hydroid to right itself is much 
longer. Two hydroids, hung vertically by a string tied to their 
proximal ends, were horizontal within seven hours, inclined up- 
ward at an angle of forty-five degrees in thirty hours, and vertical 
in their normal position in forty-eight hours. In another experi- 
ment the hydranths and hold-fasts were removed from two 
hydroids which were hung on a thread piercing them near their 
proximal ends. They righted themselves in twenty-four hours. 
When hung from strings around their necks, the stems remained 
as they were, vertical, whether they possessed hydranths or not. 
A stem lacking both hydranth and hold-fast was pierced through 
the middle by a glass needle which was suspended horizontally. 
In an hour the stem was vertical, distal end up, and remained thus 
for several days. 

Fig. I shows fairly well the important fact that it is unmistak- 
able in the animals themselves, that the stem in turning toward 
the vertical does not bend locally but generally. Corymorpha 
resembles the stems of plant seedlings in this respect, as well as 
in the preliminary bending of the stem beyond the vertical. 
Whether the bending travels progressively from oral to aboral end 

Biological Studies on Corymorpha. 407 

was not determined. There would seem to be more than an 
accidental association in these widely separated organisms of 
similar phenomena with similar structures. 

Another experiment demonstrated that a long exposure in an 
inverted position to the influence of gravity has no effect on the 
response of the individual when returned to normal conditions. 
Two hydroids were suspended vertically upside down in glass 
tubes, to prevent them from righting themselves. At the end of a 
week they were freed, and oriented themselves normally. 

There are two elements in the stem to which the foregoing 
results might be referable: the muscles, and the axial endoderm. 
To solve the problem which thus presented itself the following 
typical experiments were performed. 

A hydroid was decapitated and three wounds made at moderate 
intervals half through the stem on one side. The stem bent 
toward the opposite side, showing the greater potency of the un- 
harmed muscles. When, however, it was laid upon the bottom 
of the aquarium, wounded side uppermost, it assumed an erect 
position in about an hour. It moved toward the muscularly 
weaker side as rapidly as it would have done it had the stem been 
intact. In whatever relation to the bottom the wounds were 
placed, the stem regained a vertical position in about the same 
time — in all cases very gradually. Another individual was cut in 
a similar manner, though in this case there were eight or nine cuts 
alternately on one side and the other. These cuts interrupted 
the continuity of all the muscles except for very short distances on 
the stem, which lay quite limp on the floor of the aquarium imme- 
diately after the operation. Within two hours, however, it had 
stiffened into an erect posture. The wounds in these cases did 
not close for many hours after the stem had become erect. 

Only the continuity of the longitudinal muscles was broken by 
the wounds, whose edges were drawn apart by the contracting 
muscles. The axial cells not only maintained a continuous 
column, but bulged out into the gaping wounds in the wall, under 
considerable internal pressure. Since a stem mutilated on one 
side may right itself when it is much contracted, and in this con- 
dition the muscles as a whole on the wounded side are weaker than 

4o8 Harry Beal Torrey. 

those on the other side, it seems highly probable that the orientation 
is not the result of muscular activity, but must be due to changes in 
relative volume of the vacuolated endoderm cells on opposite sides 
of the stem. And since these cells are exceedingly large, v^ith 
excessively thin walls and almost no protoplasm, the changes in 
volume appear to be due to changes in the turgidity of the cells. 
This conclusion is borne out by the facts that the stem may not 
only shorten w^ithout increasing its diameter, but may lengthen 
yvhWe. actually increasing its diameter, results possible only through 
a variation in the turgidity of the axial cells. A complete demon- 
stration that muscles do not take part in the geotropic response 
is lacking, because in spite of the numerous transverse cuts made 
in the stem, the latter was still able to shorten (thickening at the 
same time), showing that the muscles were not rendered entirely 
impotent. But the slowness of the response and its occurrence 
while the wounds gaped and the muscles on the upper side of 
the stem were manifestly weaker than those on the lower side, 
strongly support the view that they were not concerned in the 
result. I think we may say that the muscles produce the 
movements of the tentacles, proboscis, and all save the geotropic 
movements of the stem, including shortening and possibly length- 
ening (by means of the circular muscles) while the axial cells 
cause the geotropic orientation as well as lengthening of the 

If organic growth is increase in volume,^ then the changes in 
turgidity which affect the orientation and length of the stem must 
be reckoned among growth processes, and as such they will be 
found to differ in no fundamental respect from those growth pro- 
cesses in plants and in all probability the fixed hydroids also which 
accomplish the orientation of these organisms with reference to 
gravity. This statement requires some comment. 

A comparison of the phenomena of geotropism in the stems of 
plant seedlings and Corymorpha brings out points both of resem- 
blance and difference. The cells reacting to the geotropic stimu- 

^The skeletal function of the axial cells, correlated with the alienee of a sup- 
porting perisarc, will be considered in a subsequent paper. 
^Davenport, '97. 

Biological Studies on Corymorpha. 409 

lus are in both cases strikingly similar in structure, being large, 
with a relatively small amount of protoplasm and large vacuole, 
and they bring about the bending of the stem by changing their 
volume. In the seedling, the cells on all sides of the stem increase 
in volume, but in those on the lower side the increase is greater 
than in the cells on the upper side. In Corymorpha, a similar 
differential between the upper and lower cells is established, 
though this may not involve a change in volume of cells on both 
sides of the stem. Just how it is established cannot be deter- 
mined at present, for reasons which bring about an interesting but 
not fundamental difference between the responses in plants and 
Corymorpha. The increase in volume of the plant cells — their 
growth — is permanent, because it includes growth of skeletal cell 
walls which prevent the return of the cells to their previous size, 
and it can be readily measured. The cells of Corymorpha have 
no such walls, and can change their size without difficulty; their 
growth is temporary and cannot be measured in the same way, 
because the stem is liable to frequent non-geotropic changes in 
length. The presence or absence of skeletal cell walls determines 
whether the growth is to be permanent or transitory. This con- 
sideration leads to the discussion of the geotropisms of such fixed 
hydroids as the sertularians, some of which are known to be 
geotropic, and all of which are provided with stout perisarcal 

The experiments of Driesch ('92) on species of Sertularella, 
brought out the facts that the geotropic bending of the stem is not 
general, but localized in the growing region at the end of the stem, 
and that the growth which accompanies the bending is permanent. 
My observations ('02) of Sertularia furcata and .S". argentea in 
nature are in harmony with these results. "The San Francisco 
colonies [of S. furcata] were growing on erect stalks of Phyllospa- 
dix. The stems are short and project from all sides of the eel- 
grass. Each stem leaves the eel-grass at an angle of about thirty 
degrees, then bends quickly away so that for the most part it 
makes an angle of seventy degrees with the stalk. The hydro- 
thecae of the first, and often of the second pair as well, are not in 
contact. Those of succeeding distal pairs are not only in contact 

410 Harry Beal Torrey. 

for half their length but tend much more strongly toward the upper 
side of the stem than do the proximal hydrothecae. This would 
seem to be an instance of the effect of gravity upon the direction of 
hydranth buds. The farther the stems diverge from the vertical, 
the more closely do the hydrothecae of each pair crowd each other 
on the upper side of the stem" (p. 66). "The habit of the San 
Francisco colonies of iS". argentea seems to be controlled in an 
interesting fashion by gravity. The branches are borne on all 
sides of the stems, which were fastened by their bases to the per- 
pendicular side of a shore boulder. Each stem had curved up- 
ward, so that while the basal portion was nearly horizontal, the 
terminal fourth or fifth was approximately vertical. In this 
terminal vertical portion the branches and the hydrothecae on 
them were arranged symmetrically with respect to the axis of the 
colony; and in this region the axis of the colony and the lines of 
force of gravity were parallel. At the base, where they were not 
parallel, branches and hydrothecae were oriented with respect to 
the force of gravity alone. Both hydranth and branch buds, as 
well as the stem, thus appear to be more or less negatively geo- 
tropic, the hydranths always being borne on the upper sides of 
the branches; the latter grow away from the center of the earth 
but never become parallel with the main stem" (p. 68). 

It is not difl&cult to explain why the geotropic bending is at the 
end of the stem. It is only at the growing tips of stems and 
stolons that the perisarc is dissolved. Elsewhere the cells of the 
coenosarc may contribute perisarc, but are ordinarily unable to 
dissolve it. At the tip, then, either by means of muscular activity 
{cf. the free Cerianthus) or, more probably, by growth processes 
similar to those taking place in Corymorpha — it is not yet de- 
termined which — the stem assumes an orientation which is 
temporary at first, becoming permanent only when the harden- 
ing of the perisarc about it prevents further bending. 

Enough has been said to show that no fundamental distinctions 
can be made between the geotropism of permanently fixed, tem- 
porarily fixed and permanently free organisms, such as plants, 
hydroids, anemones, protista. Any theory, then, which seeks to 
offer a thoroughly satisfactory interpretation of geotropism in one 

Biological Studies on Corymorpha. 


of these organisms, must be similarly satisfactory for all. The 
following characteristics of the geotropism of Corymorpha appear 
to find no adequate explanation in existing theories. 

Suppose a Corymorpha stem (Fig. 2) to be cut at x into two 
segments, A and B. Now, at whatever point A be supported 
above the proximal end, the latter will seek the center of the earth. 
On the other hand, at whatever point B be supported, its distal 

Fig. 2. 
Diagram of stem, to illustrate geotropism. 

end (cut at the same level as the proximal end of A) is strongly 
negatively geotropic. If the cut were made at x^, the shaded por- 
tion which, as a part of B would have been negatively geotropic, 
would now appear to be positively geotropic. I say "appear to 
be," for while the negative response is due to a change in the 
turgidity of the axial cells, the positive response may not be truly 
geotropic, but may mark an unresponsive period in the axial 


Harry Beal Torrey. 

cells, induced by a suspension of the stem from its distal end, in 
which case the stem would come to a vertical position of its own 
weight. This statement may be nearer the facts and yet not lead 
us appreciably nearer an explanation. Why should the same cell 
react in one way when the stem is attached proximally, and another 
way or not at all when the stem is attached distally ? 

In seeking an answer for this question, it should be observed 
that gravity may conceivably stimulate the stem in several ways: 
(i) through the difference in the mechanical stresses on the two 

Fig. 3. 

Diagram to illustrate geotropism. Cc, compression, greater and less; Tt, tension, 
greater and less. 

sides of the stem, (2) through the difference in the resistance en- 
countered by the organism according as it goes upward (frictions- 
weight) or downward (friction— weight) — Davenport's theory as 
applied to free organisms, (3) by redistributing the contents of the 
axial cells so that in any but a vertical position of the stem the cells 
would be in a state of unstable equilibrium with respect to the 
geotropic function. 

With reference to the first hypothesis, it may seem that the 
difference in the response of the same cell may depend upon cer- 

Biological Studies on Corymorpha. 413 

tain differences in the stresses when the stem is suspended from 
one end or the other (Fig. 3). If we assume the axis to be rigid 
to some extent, then, when the stem is anchored proximally 
(a and b) its weight may tend to compress its elements in the direc- 
tion of its axis when it is vertical (a, C); when it is not vertical (b), 
there may be added to this compression a tension of the elements 
on the upper side of the stem (Ct), and an increased compression 
of those on the lower side (Cc). When the stem is hung from its 
distal end (c and d), a tension may take the place of the com- 
pression (d, T), and if the stem be not vertical (r), the tension may 
be increased on its upper side {Tt), a degree of compression added 
to the tension on its lower side (Tc). There may be, then, a 
degree of tension on the upper side and a degree of compression on 
the lower side of each stem; in which case the differences would be 
differences of degree only. That differences of degree do not 
modify the reactions of the axial cells is evident when it is remem- 
bered that a stem hung vertically from its proximal end begins to 
right itself when, according to the hypothesis, the tension factor is 
strong on both sides of the stem, and continues in the same direc- 
tion after it has passed the horizontal, /. e., after the tension has 
ceased on the lower side and become much reduced on the upper 

The inadequacy of the first hypothesis may be shown further, 
in the discussion of the second. This view was formulated to 
explain the orientation of free organisms only. It assumes that 
negatively geotropic organisms tend to move in the direction of 
greater resistance; being heavier than water, they would meet with 
greater resistance in going upward than in going downward. 
"Another stimulus," says Davenport, "which is probably asso- 
ciated with this, depends upon the fact that an unsymmetrical 
body, heavier than water, tends to fall with its larger end down." 
That this view cannot explain the phenomena of orientation in 
Corymorpha will be clear from the following considerations: 
First, it presupposes locomotion, while locomotion is not con- 
cerned in the orientation of Corymorpha. Second, if the stem 
moves in the direction of greater resistance, a stem hung from its 
distal end ought to move in the same direction as a stem hung from 


Harry Ben I Torrey. 

its proximal end and parallel with the first (Fig. 4). Instead of 
moving in the same direction, however, they move away from 
each other as indicated by the arrows, A in the direction of less, 
B o^ greater, resistance. 

It is clear that the factor of external resistance does not govern 
such behavior, nor does the mechanical factor of tension, as has 
been shown above. It is equally difficult to explain the geo- 
tropic reactions of Corymorpha on obviously mechanical grounds 
by means of the third hypothesis; for the response of a given cell 
may be different, according as the stem is hung by its proximal 
or distal end, though the contents of the cell be distributed by 

Fig. 4. 
Diagram to illustrate geotropism. 

gravity in the same way in the two cases. This hostility to the 
familiar mechanical explanations which appear to account for 
the facts in other geotropic organisms urges upon me the desir- 
ability of repeating and extending my experiments as soon as oppor- 
tunity is afforded. It is certain, however, that the stem as a whole 
orients itself negatively to gravity, without regard to' the point at 
which it is supported. And the reactions of the axial cells are 
unquestionably associated with the polarity of the stem. A 
change in the polarity of a region of the latter is always accom- 
panied by a change in the reactions of the axial cells in this region. 

Biological Studies on Corymorpha. 415 

For instance, the cells in .v-^i (Fig. 2), if forming part of the piece 
A, would as a whole be positively geotropic so long as the aboral 
end of ^ tended to develop a hold-fast (thus preserving the original 
polarity of the piece). If, however, this end should develop a 
hydranth, the behavior of these cells would be reversed; they 
would exhibit, as a whole, negative geotropism. 

There are, then, two manifestations of polarity of apparently 
different sorts: first, polarity expressed through a special mechan- 
ism involving a single tissue, in terms of osmotic pressure and 
consequent movements of geotropic orientation ; second, polarity ex- 
pressed through a mechanism involving many tissues, in terms of 
regenerative development and differentiation. However different 
these may seem, they are undoubtedly referable to the same funda- 
mental causes — the causes of polarity in general, which involve 
internal factors at present objects of speculation only. Yet it 
may be possible to determine these internal factors more easily 
by means of the facts of what may be called functional polarity 
than by the relatively complex morphological phenomena of devel- 
opment and differentiation. The simpler mechanism and the 
simpler effects of polarity as manifested in the axial cells, are 
bound to bring a true explanation of polarity nearer our compre- 
hension, although it may still be unattainable. How an organic 
membrane or its contents may change in order to produce a change 
in osmotic pressure, while an enormously difficult problem, is yet 
more hopeful of solution than the problem of how several tissues 
simultaneously differentiate in different directions to produce a 
complicated regeneration. 

c. Locomotio72; Amoeboid Cells. 

We come now to a third type of movement, with a new cause. 
The ectoderm cells of the proximal end of the stem are capable 
of amoeboid movements, by the aid of which the hydroid may 
slowly change its location. In this regard Corymorpha closely 
resembles Hydra. On a horizontal surface, whatever locomotion 
there is takes place in any direction, with the one qualification 
that the stem moves always out of its perisarcal investment, 
which it leaves behind. 

41 6 Harry Beal Torrey. 

The rate of locomotion is slow. Half an inch in twenty-four 
hours is a maximum rate. On a vertical surface the movement is 
always directly upward. Gravity evidently determines the direction. 

The value of locomotion of this sort, and especially its negatively 
geotropic character, would seem to lie in providing a means 
whereby the hydroid may keep above the surface of the shifting 

The filaments of the hold-fast are also furnished with amoeboid 
cells by which they are enabled to move out amongst the sand 
grains to which they cling and anchor the stem {cf. amoeboid 
movements of the tips of stolons of Campanularian hydroids). 
One set of observations gave a rate of nine microns per minute at 
the tip and five microns per minute halfway to the base of a fila- 
ment several millimeters long. The free end is swollen and club- 
shaped, with well developed ectoderm which not only provides 
amoeboid cells but gland cells, which secrete the perisarc in which 
the final strength of the filament as an anchor lies. The ecto- 
derm of the remainder of each filament is attenuated almost to the 
limit of visibility; the whole filament appears to be upon the 
stretch, pulled out by the creeping club-shaped end. The endo- 
derm of the filament is composed of a single column of cells such 
as is characteristic of the endoderm of the tentacles of Campanu- 
larian hydroids. These cells, under the tension, may become 
much longer than broad, and retain these proportions whether the 
filament is attached or free. A "setting" process seems to have 
followed the stretching here, effecting the permanence of the 
attenuation without cell division. 

The direction of locomotion of the filaments is always outwardy 
but appears to be otherwise indeterminate. Arising below the 
perisarc on the peripheral canals of the stem as solid outgrowths 
with a deflection toward the proximal end, they creep along the 
stem for a short distance, closely in contact after the manner of 
stolons, and then push outward, secreting perisarc as they go. If 
the stem is hung freely in the water, the filaments extend in all 
directions. If it is in contact with the substratum, however, they 
creep along the latter as soon as they come in contact with it. 
If the substratum is sand, a filament pushes its way between the 

Biological Studies on Coryniorpha. 417 

grains, as a plant root pushes its way through the earth. In no 
case, however, does it appear to respond to the stimulus of gravity, 
or any other stimulus, except that of contact. Resistance seems 
to incite movements which overcome it. This is probably the 
reason why the filaments leave the easy path between stem and 
perisarc to push out against the resisting wall of the latter. 

d. Circulation; Cilia. 

A fourth type of motion is found in the currents set up in the 
cavities of the digestive tract by means of cilia. The cilia are 
borne on the lining cells of the proboscis of the hydroid and the 
epithelium bounding the peripheral canals on their outer side. 
They are present throughout the peduncles bearing the medusae, 
and the manubria of the latter. 

There are variations in the currents, particularly in the peri- 
pheral canals. At times there may be no current at all. At 
others the current may be setting very rapidly in the same direc- 
tion in all the canals visible. Abrupt reversals occur under these 
conditions, which can hardly be explained by ciliary action, but 
are rather the result of expansions or contractions of the proboscis 
and stem, which produce changes of pressure in the canals. 


The eggs are laid by medusae which are never set free from 
the hydroid. They are small but heavy with yolk and fall directly 
to the bottom in quiet water, adhering by their delicate coats to 
the first object they touch. 

As soon as the egg is attached, its free life is practically over. 
The embryo is never ciliated^ and has no free-swimming phase in 
its existence. It is capable of very slow creeping movem'ents, 
however, by means of which it often comes in contact with other 
embryos and forms with them temporary associations of as many 
as six, ten, twelve individuals. Often it will travel many times its 
own length, leaving behind a narrow collapsed tube of perisarc 

iC/. Hypolytus peregrinus, Murbach (Q. J. M. S., XLII, 1899, p. 341), which it 
resembles in this and other respects. 


Harry Beal Torrey. 

which it has secreted and which is continuous with the egg case. 
As we have seen, the hydroid never loses its power of locomotion, 
even after the development of the filaments of the hold-fast. 

As the embryo leaves its egg case, it elongates, and an anterior 
(oral) end can be distinguished from the narrower posterior 
(aboral, proximal) end. The anterior end soon elevates itself, 
and the embryo now touches the substratum by one side of the 
aboral region only. 

For about thirty hours, or up to the time when the hydranth is 
beginning to form, the embryo is completely covered by an 

Fig. 5. 
Young Corymorpha. 

extremely delicate layer of perisarc. From this time the perisarc 
is frequently limited entirely to the stem. Before the formation 
of the hydranth, the perisarc covering the anterior end of the 
embryo, and secreted by glandular cells of the ectoderm, is not 
permanent, being dissolved as the stem progresses, probably by 
the secretion of other cells in this region. As the hydranth begins 
to develop, its ectoderm ceases to manufacture perisarc, which 
henceforth is deposited by cells beginning at the aboral limit of 

Biological Studies on Corymorpha. 419 

the hydranth. The perisarc is hardly sturdy enough at any time 
to afford any support to the stem. Its adhesive character, how- 
ever, serves to attach a portion of the latter to the substratum, over 
which the coenosarc creeps. 

Amoeboid ectoderm cells are responsible for the locomotion of 
the young Corymorpha (Fig. 5) as of the adult, though they are 
not confined to the proximal end of the stem. Often the latter 
clings for half its length and may perform looping movements, 
much less pronounced, however, than those of Hydra} The 
direction of locomotion is also determined by the same factors 
which regulate it in the adult. On a horizontal surface the direc- 
tion is indeterminate, though the stem always moves out of its 
investment. On an oblique surface it tends to move upward by 
the nearest route. Young hydroids are often found adhering to 
the stem of an adult, the relation of the axis of the attached por- 
tion of each to the adult axis varying with the inclination of the 
latter. If it is vertical, they are parallel with it, vertical also; and 
the rest of the young stem will be nearly vertical, but not quite so 
since the distal portion of the stem seems to shun any contact 
(negative thigmotropism). If the orientation of the adult is 
altered, the young hydroid will gradually take up a new position 
in which the most distal point of attachment will be the greatest 
possible distance above its proximal end. 

Not only, therefore, is the larva negatively geotropic with regard 
to orientation, but this has a directive effect upon locomotion. 
It is probably due to the effect of the stimulus of gravity on 
the endoderm cells which line the single cavity and from which 
the axial cells of the adult are derived. These cells do not con- 
tain such enormous vacuoles as those in the axial endoderm, and 
are ciliated. In these respects they resemble the parietal endo- 
derm cells of the peripheral canals of the adult, which are their 
descendants also. 

With reference to the amoeboid cells which produce locomotion 
in Corymorpha, it may be recalled that the ectoderm of hydroids 
is not uncommonly amoeboid. To cite but a single instance, not 

'C/. MarshaU. Zeitschr. f. w. Zool., XXXVII, p. 664. 

420 Harry Beat Torrey. 

only is the cauline coenosarc in Campanularian hydroids, e. g., 
Obelia, fastened to the perisarcal tube here and there by multi- 
cellular amoeboid processes of the ectoderm, but the anterior ends 
of growing stolons exhibit amoeboid changes of form which ac- 
count for their creeping movements and produce the tension often 
manifested in the coenosarc which is fixed farther back on the 
stem. The coenosarc is literally dragged out of the perisarc. 
A similar tension has already been noted in the filaments of the 
hold-fast, due to a similar cause. And it is probable that the 
proximal end of the larva may be dragged along at times after the 
more distal attached portion. 

The active muscular movements discussed at length for the 
adult need not be considered here, as the young hydroid appears 
to respond similarly in all respects, with the one exception that the 
reaction times are somewhat greater. 

The absence of a free-swimming larval form seems to account 
for the tendency of Corymorpha palma to dwell in communities, 
as previously mentioned. The power of locomotion is too slight 
to have any effect on the distribution of individuals, which is 
accomplished by tidal currents and the shifting of surface sands. 
Occasionally an individual may be washed away from its anchor- 
age, and begin a new community in a new locality. 


Corymorpha is unusually active for a hydroid. It is every- 
where sensitive to mechanical stimuli, irritating chemicals and 
abrupt changes in temperature, nowhere to "odorous" substances. 
The prehensile mechanism is composed of proximal tentacles, 
which move toward the mouth in response to all eff'ective stimuli; 
distal tentacles, which move away from the mouth in their initial 
response to stimuli; and proboscis, which may move toward the 
point stimulated. These movements, as well as shortening and 
possibly lengthening the stem, are performed by muscles. 

The stem of the adult responds to the stimulus of gravity, by 
means of a change in the turgidity of the vacuolated axial cells. 
The response of these cells varies according as the stem is attached 

Biological Studies on Corymorpha. 421 

proximally or distally, and according as it is heteromorphic or not. 
The polarity of the stem is expressed, not only by the regenerative 
development but by the changes in the axial cells. 

Locomotion is accomplished by amoeboid cells located at the 
proximal end in the adult, more generally distributed in the larva, 
and covering the club-shaped ends of the filaments of the 

Cilia are present on the epithelial cells lining the hydranth 
cavity and peripheral canals. Supplemented by contractions and 
expansions of the hydranth cavity, they provide for the circula- 
tory currents through the body. 

Eggs are laid both in summer and winter, usually during the 
morning hours. They have adhesive coats. The planulae are 
never ciliated, and their locomotion is limited to very slow creep- 
ing movements. The larvae are geotropic. 


Agassiz, a., 1862. — Contributions to the Natural History of the United States. 

Vol. IV, p. 276. 
Allman, G. J., 1863. — Notes on the Hydroida. Structure of Corymorpha nutans. 

Ann. and Mag. Nat. Hist., 1863, p. 1. 
1871. — A Monograph of the Gymnoblastic or Tubularian Hydroids. 

CzAPEK, Fr., 1898. — Weitere Beitrage zur Kenntniss der geotropischen Reizbewe- 

gungen. Jahrb. wiss, Bot., Vol. XXII, p. 175. 
Davenport, C. B., 1897-9. — Experimental Morphology. Parts I and II. New- 
Davenport and Perkins, 1897. — A Contribution to the Study of Geotaxis in the 

Higher Animals. Jour. Physiol., Vol. XXI, p. 22. 
Driesch, H., 1892. — Kritische Erorterungen neuerer Beitrage zur theoretischen 

Morphologie. II. Zur Heteromorphose der Hydroidpolypen. Biol. 

Centr., Vol. XII, No. 18, p. 545. 
EiMER, Th., 1878. — Die Medusen, etc. Tubingen. 
Haberlandt, G., 1901. — Ueber Reizleitung im Pflanzenreich. Biol. Centr., Vol. 

XI, p. 369. 
Holmes, S. J., 1904.— The Problem of Form Regulation. Arch. Entmckl., Bd. 

XVII, p. 265. 
Jennings, H. S., 1902. — Studies on Reactions to Stimuli in Unicellular Organisms. 

IX. On the Behavior of Fixed Infusoria (Stentor and Vorticella) 

with Special Reference to the Modifiability of Protozoan Reactions 

Am. Jour. Physiol., Vol. VIII, p. 23. 
1904. — Contributions to the Study of the Behavior of Lower Organisms. 

Carnegie Institution of Washington. Publ. No. 16. 

422 Harry Beal Torrey. 

LoEB, J., 1891. — Ueber Geotropismus bei Thieren. Arch. f. ges. Physiol., Bd. XLIX. 
1891a. — Untersuchungen zur physiologischen Moiphologie der Thiere. I. 

Ueber Heteromorphose. Wiirzburg. 
1892. — Untersuchungen zur physiologischen Morphologic der Thiere. II. 

Organbildung und Wachsthum. Wiirzburg. 
1894. — On Some Facts and Principles of Physiological Morphology. Woods 

Hole Biol. Lect. 
1895.- — Zur Physiologic und Psychologic der Actinien. Arch. f. ges. 

Physiol., Bd. LIX, p. 415. 
1902. — Comparative Physiology of the Brain and Comparative Psychology. 
New York. 
'Mast, S. O., 1903. — Reactions to Temperature Changes in Spirillum, Hydra, and 

Fresh-Water Planarians. Am. Jour. Physiol., Vol. X, p. 165. 
May, a. J., 1903. — A Contribution to the Morphology and Development of Cory- 

morpha pendula Ag. Am. Nat., Vol. XXXVII, No. 441. 
Morgan, T. H., 1901. — Regeneration. New York. 

1901a. — Regeneration in Antennularia. Biol. Bull., Vol. II, No. 6. 
MuRBACH, L., 1899. — Hydroids from Woods Hole, Mass., etc. Q. J. M. S., n. s., 

Vol. XLII, p. 341. 
Nagel, W., 1893. — Versuche zur Sinnesphysiologie von Beroe ovata und Carmarina 
hastata. Arch. f. ges. Physiol., Bd. LIV. 
1894. — Vergleichend physiologische und anatomische Untersuchungen 
ueber den Geruchs-und Geschmackssinn und ihre Organe. Bibl. 
Zool., Heft 18. 
1894a. — Experimentelle sinnesphysiologische Untersuchungen an CcBlen- 
teraten. Arch. f. ges. Physiol., Bd. LVII, p. 495. 
Nemec, B., 1901. — Ueber die Art der Wahrnehmung der Schwerkraftreizes bei den 

Pfianzen. Ber. deut. Bot. Ges., Bd. XVIII, p. 241. 
Parker, G. H., 1896. — The Reactions of Metridium to Food and Other Substances. 

Bull. Mus. Comp. Zool., Vol. XXIX, p. 107. 
Platt, J., 1899. — On the Specific Gravity of Spirostomum, Paramoecium and the 
Tadpole in Relation to the Problem of Geo taxis. Am. Nat., Vol. 
XXXIII, p. 31. 
Romanes, G. J., 1876. — Preliminary Observations on the Locomotor System of 
Medusa;. Phil. Trans. Roy. Soc. Lond., Vol. CLXVI. 
1877. — Further Observations on the Locomotor System of Medusse. Ihid., 
Sars, M., 1853. — On the Nurse Genus Corymorpha and its Species, together with the 
Medusse Produced from Them. Ann. and Mag. Nat. Hist., Vol. 
VIII, p. 353. 
Torrey, H. B., 1902. — The Hydroida of the Pacific Coast of North America. Un. 
of Cal. Publ. Zool., Vol. I, p. 1. 
1902a.— American Naturalist, Vol. XXXVI, p. 987. 

1904. — The Habits and Reactions of Sagartia davisi. Biol. BuU., Vol. VI, 
No. 5, p. 203. 
True, R. H., 1895. — On the Influence of Sudden Changes of Turgor and of Tempera- 
ture on Growth. Ann. of Bot., Vol. IX, p. 365. 
Verworn, M., 1899. — General Physiology. Trans, by F. S. Lee. New York. 





With 3 Plates and 3 Figures in the Text. 

Experiments on the life-history of Paramcecium caudatum have 
now been carried on continuously for 29 months. Two series, 
designated as the "A series" and the "B series," were started on 
the first of February, 1901, with individuals from different 
sources. The B series died out in May, 1902, in the 570th genera- 
tion; the A series on December 19, 1902, in the 742d genera- 
tion. A third series — "C" was started in June, 1902, with an 
individual from Cambridge, Mass., and died out in June, 1903, in 
the 379th generation. The progress of the first two series has 
been recorded from time to time,^ and in the present paper I 
wish to give the history of the last cycle of the A series and to 
consider the results in relation to some general biological prob- 
lems and theories. 


As described in the earlier Studies (I and III) the general 
vitality of the two series, A and B, as expressed by the daily 
division rate, underwent periodic cycles of vigor and depression. 

^(1) Studies on the Life History of Protozoa. I. The Life Cycle of Paramcecium 
caudatum. Archiv. f. Entwk. XV, 1, 1902. 

(2) Studies, etc. II. The Effect of Stimuli on the Life Cycle of Paramcecium 
caudatum. (With C. C. Lieb). Arch. f. Protistenkunde. I, 1, 1902. 

(3) Studies, etc. III. The 620th Generation of Param. caud. Biol. Bull. Ill, 5, 

424 Gary N. Calkins. 

The early curves appeared to indicate a periodicity of three- 
month intervals, and this v^as taken to be the time of the usual 
life cycle in culture of Paramcecium candatum; this conclusion w^as 
based partly upon my ow^n results and partly on those of Jou- 
kowsky and of Simpson, both of whom found that cultures of this 
infusorian died out after three months of treatment. It w^as 
found, however, and it may be seen from the now completed 
curve of the A series (see Diagram I) that trimonthly periods of 
depression were not fatal and that recovery occurred without 
purposeful stimulation. Thus in the first apparent depression 
(May, 1901,) the recovery was thought to be due to the stimula- 
tion by jolting on a railroad trip of six hours; another in March, 
1902, was considered due to a slight rise in temperature. These 
periods of depression differ markedly from those of August and 
December, 1901, and of June, 1902, when the individuals con- 
tinued to die at a high rate, notwithstanding repeated jolting 
experiments, increase in temperature, and the like, and the race 
was saved only by change to a special diet after numerous 
attempts and failures with foods of different kinds. The well- 
marked cycles, therefore, with periods of depression which de- 
manded stimulation of a decided character, were approximately 
of six months' duration, while intermediate cycles of less impor- 
tance were about three months long. The first of the six-month 
cycles ran from February i, i90i,to August i, 1901, (see Diagram 
I); the second from August 15 to January i, 1902; the third from 
January i to July i, and the last from July to December 19, 1902. 
During the first three cycles the number of generations was nearly 
the same (200, 198 and 193, respectively), the last, on the other 
hand, was much less, the individuals dividing only 126 times. 

The stimulation which resulted in the renewal of vitality after 
the periods of depression in August and December, 1901, was due 
to the change from hay infusion diet to beef extract for a 
limited period (see Studies I and III). The same change failed to 
work in the July, 1902, period of depression, and after the race 
had become reduced to only six individuals, a successful sub- 
stitute for the beef extract was found in the extracts of pancreas 
and brain (see Studies III). Recovery, however, was not so 

Studies on the Life History of Protozoa. 425 

successful as in the previous periods and the organisms were much 
less vigorous than at similar periods in previous recoveries. The 
division rate, furthermore, slowly fell from the relatively high 
point in August, and gradually decreased during the fall months 
until the A series died on the 19th of December. The B series 
had succumbed in the 570th generation, in June, before the right 
stimulus was found. Except for the slowness of divisions the 
organisms appeared perfectly healthy during the summer and fall 
of 1902, although microscopical study of preparations made during 
this period showed characteristic changes in the protoplasmic 
structure (see Figs. 18 to 21). The organisms were plump and 
moved freely about the slide, responding with customary vigor to 
stimuli of diflperent kinds. Every precaution was taken during 
this period to invigorate the race and every experiment that my 
ingenuity could devise wks executed; some appeared to give a 
temporary improvement but none was permanent, and the last 
individual of the A series finally died after 23 months of continued 
daily observation, without, however, any morphological evidence 
of general senility. (See Diagram I.) 


Artificial rejuvenation of the A series was successfully accom- 
plished three times. The experiments and results have been 
described in other places, and it will be sufficient here to merely 
point out that after considerable experiment, beef extract was 
successful in the first two cases and pancreas and brain extract in 
the third, the result being due, probably, to the change in salt 
contents of the medium. The approaching end of the series was 
indicated some time in advance by the reduced division rate 
during the fall of 1902, and efforts were continuously made to 
rejuvenate them during this period. For these experiments all of 
the stock of the regular series was maintained, and the number of 
lines under observation frequently ran up to twelve or more. 
The results of all experiments were tabulated and the effects of 
the stimuli used were noted for comparison with the regular 
series. The general result may be seen upon the diagram which 


Gary N. Calkins. 


n '■'< 

% r 

--', 'V-" " 

'"■; " 1 " 


..'. .:::^::i.. 


.._!.'.. _:l _..: 

....- r 

' [-:■ 

M ___i.... 

is ^— ].:._.. 

i ^....__,.. 


""1 :::}EB 





CI -' 

O ^ 

a> o 


-5= r^ 

_ o 

Oi ^ 
1^ t-i 


S .2 

o be 

a a 














^ '2 
S -S 


05 03 •-' 




























O .5 
tin H 


Studies on the Life History of Protozoa. 427 

hows that, despite all efforts to stimulate, the race rapidly 
weakened and ultimately died. 

I. Experiments with Extracts. 

a. Beef Extract. During the last autumnal period, beef extract 
was used at different times in the same way that it had been used 
successfully during previous periods of depression, and for the same 
length of time, twenty-four hours. The organisms were immersed 
in the fluid full strength in the majority of cases, but experiments 
with the half strength were also made. The failure of the beef 
treatment in June, 1902, has already been described (Studies III) 
and I shall consider in this place only the experiments subsequent 
to that time. In general it may be stated that beginning with* 
the treatment in May, the effect of the beef extract was nil. On 
the 19th of June A3 and A4 were immersed as usual and for the 
same length of time. Both of them died before the 27th. Again 
on the 22d Ai and A2 were treated, and both of these died on 
the 26th. Similar results were obtained in all later experiments, 
as shown by the following resume: 

Both died before the 15th. 

Ai and A2 died on 25th. A5 on 

the 26th. 
Both died before the 19th. 
Both died before the 13th. 
Died on the 27th. 
Died on the 29th. 
Died December 2. 

The short time in beef may not have been long enough to make 
the change beneficial; with this in mind I kept the last few 
individuals in for three days (Nov. 26, 27, 28). Not one of them 
divided more than once and all died within a week. Beef extract, 
therefore, had lost its potency as a rejuvenating medium. 

The effect of beef extract upon the body structures was to 
increase the number of gastric vacuoles; while, in some cases, the 
micronuclei were caused to divide (Figs. 5 and 6). Even in May, 
1902, there was an indication of the endoplasmic concentration 

July. 6. 

A3 and 

A4 treated 

July 24. 

Ai, A2, 

A5 " 

Aug. 2. 

Ai and A2 " 

Aug. 2. 

A5 and A6 " 

Nov. 26. 



Nov. 26. 



Nov. 26. 



Nov. 28. 


Aq and Ai: 

428 Gary N. Calkins. 

which accompanied depression at this period. The dense con- 
dition of the protoplasm is better shown in Fig. 6 which represents 
an individual twenty-four hours after transference from beef 
extract into hay infusion. It may be noted here that, at this 
period, the beef extract failed to reduce the dense endoplasmic 
condition to one of tenuity which seems to be the normal 

b. Extract of Pancreas. Extract of pancreas was made in the 
same way as the beef extract. A fresh sheep's pancreas was Gut 
in small pieces and brought to boiling point in water. After 
filtering and cooling, the Paramoecia were placed in it and left for 
24 hours, as in the beef. At first it proved a good substitute for 
the beef and the organisms appeared to thrive on it; but later, in 
November and December, it was as useless as the beef. The 
following records show this fact: 

June 27. A3 treated. Divided twice in 24 hours. Forms 

the regular series from this time. 

June 29. Ai and A2 " Divided a few times. Died out 

on July 14. 

July 16. A4 " Died the next day without divi- 


July 17. A6 " Divided twice the next day; lived. 

July 18. Ai,A2 and A4 " All died the next day. 

July 20. Ai,A2andA4" Lived. Given mutton broth on 

23d. Died on 24th. 

Aug. 20. A2 " Died on 28th. 

Aug. 20. A8 " Died on the 2ist. 

Dec. 8. A5, A6, A7 and A8 treated. No divisions. All died out 

on 19th before or after treatment with various other substances. 

At the period in June when recovery was effected by using the 
extract of pancreas, the organisms of both series were in the con- 
dition represented by Fig. 7. The endoplasm was densely granular 
and homogeneous, and had a curiously "stuffed" appearance. 
This condition was relieved by using extract of pancreas, whereas 
beef extract, made with the same water and in the same way, was 
ineffectual. Figs. 8, 9 and 10 show the general course of the 

Studies on the Life History of Protozoa. 429 

action of the pancreas extract. Fig. 8 represents an individual 
twenty-four hours after treatment, i. e., after change from pancreas 
extract into hav infusion.^ The characteristic dense structure is 
distinctly shown, but in the center there is unmistakable evidence 
of the normal condition. Figs. 9 and 10 represent two indi- 
viduals forty-eight hours after treatment with the pancreas extract. 
In the former, the characteristic dense structure is still visible at 
the two ends, but the center is clearing. In the latter, new gastric 
vacuoles have appeared in the endoplasm, the animal being well 
on toward recovery when killed. 

c. Extract of Sheep's Brain. This was made in the same way 
as the other meat extracts, and the animals were similarly treated 
with it. It was not efficient as a permanent stimulant, and was 
discarded in subsequent treatment. 

d. Extract of Mutton. '' Mutton Broth.'' This extract was also 
tried in the summer (July 20 and 23), but in no case was it 
successful, the organisms invariably dying within 24 hours. 

e. Lecithin. A trace of pure lecithin was put into the regular 
hay infusion during the week of August 20. The organisms were 
apparently not injured by the change, but did not live more than 
48 hours after the treatment. 

/. Pineapple Extract. With the view of ascertaining if some 
of the vegetable ferments might not prove beneficial, I tried 
extract (juice) of fresh pineapple, and of fresh apple. A4 was 
put into dilute pineapple juice July 27. The reaction was well 
marked, as shown by decided increase of movement and by three 
divisions in the ensuing 48 hours. The experiment was repeated 
the next day with a like result. It was repeated again August 3, 
but was unsuccessful, the organisms dying two days after treat- 
ment. The stimulation was temporary in all cases, and it should 
be noted that the organisms were in a period of increasing vitality 
when the first pineapple treatment was given (see Diagram I). 

^ The hay infusion was made every day, the same amount of hay and water being 
taken each time and raised to the boiUng point. This method was never varied 
during the entire period of the cultures and the salt content of the water, as shown 
by weekly analyses, did not vary beyond a very slight fraction of one part to one 
hundred thousand. 

430 Gary N. Calkins. 

g. Apple Juice. A piece of fresh Porter apple was allowed to 
lie for a few minutes in the hay infusion. In this case the result 
was well marked, and a decided stimulus was noted. Again, on 
Sept. 20, A5, A6, A7 and A8 were all put into one drop of apple 
juice to 12 drops of hay infusion and left for thirty to forty-five 
minutes. They were then transferred to clear hay infusion and 
left. All divided the next day. The experiment was repeated on 
the 2 1st with a like result. In some cases the organisms died 
immediately, showing that the strength used was too great. When 
properly diluted, however, apple seemed to give a satisfactory 
temporary stimulus, although in no case did the stimulation last 
for more than forty-eight hours. The same experiment tried in 
October gave no results; the organisms died. 

In addition to the above, various proprietary mixtures were 
tried from time to time. Among these were phospho-albumin, 
and nuclein-albumin; none gave satisfactory results. 

2. Experiments with Acids and Salts. 

In view of the successful results which have followed experiments 
with ions in connection with egg development, it was thought that 
perhaps dilute acids or salts would have a beneficial result in the 
case of these weakened infusoria. Normal solutions were made in 
each case and various strengths were tested from those that would 
kill to those that only slightly stimulated. The organisms were left 
in the fluids for only a short time (20 to 30 minutes) and were then 
transferred to fresh hay infusion. Attention may be called here 
to the fact that potassium phosphate when used in this way was 
successful in restoring the vitality of weakened Paramcecium in 
the preceding December cycle, the "rejuvenation" which resulted 
was directly comparable with that eff^ected by the beef extract. 
There was reason, therefore, to believe that the repeated use of 
various salts would give satisfactory results in the last period of 
weakness of the race. This expectation, however, was not real- 
ized for none of the chemicals used in the fall and winter of 1902 
was successful in this way; all were as futile as the beef and pan- 
creas extract, as shown by the following experiments: 

Studies on the Life History of Protozoa. 431 

A. Potassium Salts. 

a. K^HPO^. On the 8th of June, 1902, one individual from 
the line of A2 was treated for 30 minutes with a solution of dibasic 
phosphate of the strength of one drop of ^^^n to six drops of usual 
hay infusion. The result was a marked increase in the rate of 
division for a considerable period as compared with the control 
series, as follows: 

Average daily division-rate for 5-day periods, June 8-July 5. 

Stimulated Aj. Control Series. 

1st Period 8 8 

2d *' 8 6 

3d " 8 6 

4th " i.o 2 

5th " 1.6 o 

On the 27th of June the above experimental series was 
substituted in the regular culture series and the descendants 
of these individuals formed the regular lines until the final 
extinction, subject, of course, to the other experiments as stated 

A stronger solution (1-5) and for a longer period (i hour) was 
used with A3 on June 26. The individual died in three hours. 
On August 2 1 the same strength was used, but for only 25 minutes; 
the individual died in four days without dividing. The general 
effect of this salt was, therefore, favorable, with evidence that a 
certain optimum strength is alone beneficial. The beneficial effects 
upon the endoplasmic structure are shown in Figs, il and 13. 

b. KH^PO^. Experiments with the monobasic salt were also 
made and various proportions were used, but none was successful. 

c. KCN. Various proportions of "0 of this salt were used, the 
most successful being one drop of the solution to twelve drops of 
hay infusion. This was not strong enough to kill the bacteria 
which afford the only food for Paramcecium. Four individuals 
were immersed October 29 in the mixture and left for 24 hours. At 
the end of this time each had divided once, while none of the con- 
trol series had divided. Of the eight individuals resulting from 
this treatment, four were placed again in the KCN solution, 

432 Gary N. Calkins. 

(made fresh), while the other four were placed in hay infusion 
without the salt. Of the former set, each individual divided once 
in 48 hours, and of the latter set, one died, two divided once, and 
one divided twice in the same period. The regular control series 
did not divide at all during this time. Both sets were placed in 
hay infusion on the fourth day and neither set continued to live, 
all died before the sixth of November. Another set of four in- 
dividuals were treated with the salt every day for the same period. 
After the first 24 hours none had divided; after the second 24 
hours each one had divided once. This was November 2. On 
the 4th all had divided twice, on the 6th only one had divided 
again, on the 7th another one had divided, on the 8th none had 
divided again, on the 9th one died, while the rest did not divide, 
on the loth two others died, and the remaining one was placed in 
the usual hay infusion without the salt, having been treated 
daily for ten days with it. It did not divide again until the i8th, 
and finally died out on the 21st. Others, however, that had been 
treated, and had been placed earlier in hay, continued to live and 
supplied the regular lines of the experiment. The use of the 
KCN therefore can be said to have been successful to a limited 
extent, and, possibly, to have prevented an earlier extinction of 
the race. The effect on the curve of the life cycle is shown by the 
temporary rise during the last period in October and the first 
period in November. 

d. KOH. This was tried only once with four individuals on 
the 28th of October. ^^^-^ i part to 4 was used for 30 minutes. 
On the first of December two had died and one had divided once. 
None divided again and all of the individuals experimented with 
died before the fourth of November. 

B. Sodium Salts. 

a. Dibasic Sodium Phosphate. Three individuals, Al, A3 
and A4 were placed for 30 minutes in ^S^^ NajHPO^. All 
died without dividing by the 27th. The experiment was not 

b. Sodium Tartrate. July 10 two individuals were placed for 
30 minutes in -^^ sodium tartrate, one drop to five of hay infusion. 

Studies on the Life History of Protozoa. 433 

They died in twenty-four hours. The experiment was repeated 
on July 14, three drops to five of hay infusion being used. On 
the 15th they did not divide, on the 17th they divided once, and 
died on the i8th. Experiment not repeated. 

c. Sodium Chloride. This salt was used on several occasions with 
negative results as a rule. (See, however, table below.) In 
September, 1902, when the race was comparatively vigorous, an 
individual was treated for 30 minutes with f^ NaCL, one drop of 
the salt to twelve of the hay infusion. At the end of 48 hours it 
had divided once, but died within five days without further 
division. The effect upon the protoplasmic structure was not 
particularly noticeable (see Fig. 18). 

The following table gives a comparative view of the efficiency 
of different salts on the division rate for thirty days subsequent to 
treatment. Several individuals of the A series were treated on 
the 20th of March with potassium phosphate and the progeny of 
one of these in the 78th generation were again treated in part on 
May 6 with potassium phosphate, and in part, with potas- 
sium chloride, magnesium chloride, sodium chloride and calcium 
chloride with the strengths, and for the times indicated. The 
following notes were made at the time of the treatment. "When 
the individual was put into the potassium chloride it began at 
once to swim backward with great rapidity, and continued this 
for about five minutes. It then straightened out and appeared 
perfectly normal in the solution. When returned to the hay in- 
fusion at the end of the treatment, it went through the same con- 
vulsions but soon became normal, perhaps slightly swollen and 
transparent." Again: "When the individual (another individual 
of course) was put into the magnesium chloride solution it was 
hardly affected in any way, a very slight increase in movement 
being noticed." Again: "Treatment with NaCl did not affect the 
individual, it appears fat and happy in the hay-infusion." Again: 
"Very much affected by the CaCl2 solution. One of the three 
specimens died; the other two were distorted and badly shrunken, 
this lasted for at least fifteen minutes after they had been trans- 
ferred to the hay infusion." 

434 Gary N. Calkins. 

Average number of divisions per day after stimulation. 

K.HPO^. KCI. MgCU. CaCl,. NaCl. 

May 6-10 1.60 1. 00 1. 00 0.60 1.20 

May 11-15 1. 00 1.20 0.60 1.20 1.20 

May 16-21 1.50 1.50 1. 00 1.50 1.66 

May 22-26 1.80 2.00 1.80 1.80 2.00 

May 27-June I . . . . 1.25 1.03 1.25 0.75 1.50 

June2-6 1.33 1.33 1.33 0.16 1.33 

June 7-12 0.20 0.40 1.60 dead 0.40 

dead dead living dead 

All were normal solutions, diluted 25 times, one drop to twelve 
drops of hay infusion and the treatment lasted for 30 minutes in 
case of KCI, and for 25 minutes in each of the other solutions. 
Definite conclusions cannot be drawn from one set of compari- 
sons for it may have been pure accident that the magnesium 
chloride specimens continued to live. The effect of MgClj upon 
the protoplasmic structures is shown in Fig. 12. 

Comparatively few experiments were made with acids. Hydro- 
chloric and nitric acids were tried during the period of depression 
in October, 1902, but the results were negative, the individuals 
dying within twenty-four to forty-eight hours. An interesting 
effect was produced by treatment with dilute phosphoric acid. 
The dense endoplasm was broken up and with it the macro- 
nucleus which, after the treatment, appeared as many small 
fragments (see Fig. 16). 

Of the other unsuccessful attempts to rejuvenate the race during 
the last period of depression I will mention only those with gal- 
vanic stimuli, with nitro-glycerine, and with dried and powdered 
Paramoecium of an entirely foreign race. 

3. Galvanic Stimuli. 

A small cell was made and connected with two Mesco batteries. 
Four individuals were treated on November 28, three different 
times to the full current and for a period of one minute each time. 
The usual reaction followed the treatment, migration to the nega- 
tive pole, and when the current was reversed, migration from the 

Studies on the Life History of Protozoa. 435 

positive to the negative pole. At the end of the treatments the 
four individuals appeared normal. On the following day one had 
died, another on the ensuing day, and the last two on the fourth 
day. Another time the same experiment was tried but with only 
one minute of exposure. The result was the same, death without 
division. The death of these organisms at this time cannot neces- 
sarily be ascribed to the treatment, for a glance at the diagram 
shows that the entire race was dying and that divisions were infre- 
quent in all cases. 

4. Nitro-Glycerine. 

At the suggestion of Professor Wilson, and as a last resort, I 
tried two experiments when the race appeared to be dying out in 
December. Nitro-glycerine in very weak solution (unfortunately 
I have no record of the strength used) was put into the hay 
infusion. It made no appreciable difference in the final result 
and the organisms did not divide. 

Professor Wilson's other suggestion seemed more hopeful, on 
the a priori ground that renewal of vitality is effected by the union 
of two individuals. A culture of Paramcecium fresh from 
pond water was made, and hundreds of individuals were allowed 
to dry in a small drop of water in a watch crystal. When dried 
the remains were scraped together and pulverized, the powder 
thus formed being added to the hay infusion in which the weak- 
ened Paramcecium were kept. Although this extremely ingenious 
suggestion was worthy of a fruitful result, the outcome of the 
experiments was the same as with all the rest, and not a single 
individual lived after the 19th of December, one week after six- 
day treatment with the dried Paramcecium. 

There remain many experiments that might have been tried, 
and that might possibly have accomplished the same results that 
were obtained in the earlier periods of depression when the race 
was successfully reinvigorated by artificial means, and even the 
experiments that were tried might have been successful if different 
strengths, or times of action, had been used. Many suggestions 
were made by my colleagues and other friends, especially in 
regard to the trial of some chemical compound. I am pleased to 

436 Gary N. Calkins. 

acknowledge the friendly and scientific interest which prompted 
these suggestions, and desire to state that if they were not always 
carried out, it was because of the limits of my time and of the con- 
stantly decreasing number of individuals left to experiment with. 
It was my desire to try as many classes of experiments as possible, 
and some of these might have been successful if tried at an earlier 
time or if carried out on a sufficiently large scale, but here again 
the scarcity of living material would not allow the continued 
experimentation along lines that were fruitless on the first trial. 
It must be remembered that such experiments, to be of any value 
in a work like this, had to be made on the material that had been 
under constant observation for nearly two years, and preliminary 
experiments with fresh forms from the ponds were valueless so 
far as indicating the effect on the vitality of the race under 

I. The Norrnal Paramceciutn. 

The usual size of a normal Paramcecium is from 200 to 300 
microns, and the form is fairly constant, warranting the designa- 
tion "slipper animal." In all of the preserved specimens that I 
have made from time to time, the fixing fluid was saturated cor- 
rosive sublimate to which was added 10 per cent of glacial acetic 
acid. Having a common method of fixation the different indi- 
viduals can be compared point by point. 

a. The Endoplasm. The endoplasm of a normal form is made 
up of various granules of diff'erent sizes, of vacuoles and crystals 
(Fig. i). When the animal is moving about in a nutrient 
medium it constantly takes in food with the water absorbed. 
The food of Paramcecium consists of bacteria, and these accumu- 
late in a gastric vacuole until the latter has attained a certain size 
when, according to Wallengren,^ it is caught up in the endo- 
plasmic flow and carried to the posterior end of the body. It then 
moves anteriorly toward the left side, ultimately passing over to 
the right and then down on the right side. In this migration of 

^ H. Wallengren. Inanitionserscheinungen der Zelle. Zeit. f. Allg. Physiologie 
I.. 1, 1901. 

Studies on the Life History of Protozoa. 437 

the vacuole the food is brought into the immediate vicinity of the 
macronucleus where the effect of the nuclear environment is 
shown by the immediate acid reactions with congo-red of the 
vacuole contents (Wallengren). The food material in such a 
vacuole is massed into a more or less homogeneous body corre- 
sponding to Greenwood's observation on Carchesium, and in this 
condition the digestive fluids work upon it to resolve it into 
digestible and indigestible parts. After this the soluble portions 
are absorbed and the residue defecated. The soluble portions 
pass into the endoplasm to be stored up as reserve food (Wal- 
lengren) from which they are taken as the need comes to be 
made into living molecules. 

The processes of digestion thus given rise to definite elements 
in the endoplasm, elements which react to stains in characteristic 
ways. In addition to these, however, we might expect to find 
waste matters due to incomplete oxidation as well as final products 
of metabolism in the form of crystals, etc. The various possibili- 
ties of this nature have given rise to different interpretations upon 
which my own observations throw but little additional light. 
With neutral-red acting upon the organism when alive, Prowazek^ 
distinguished three kinds of granules in the endoplasm: (i) The 
food balls; (2) Small round granules which are distributed 
throughout the periphery more or less uniformly (Prowazek 
actually found them at the two extremities and about the mouth); 
(3) Minute granules distributed throughout the endoplasm and 
all over the body. 

The minute peripheral granules (No. 2) are interpreted by 
Prowazek in the same way that Wallengren had previously inter- 
preted similar bodies in Pleurocoptes hydractinicB, viz., as excre- 
tory vacuoles with a solidified granule of excreta within them. 
Piitter,2 on the other hand, interpreted them as basal bodies of 
centrosome nature lying at the bases of cilia. Wallengren subse- 
quently showed, however, that the granules in question are not 
at the bases of cilia but lie beside the cilia, and that rows of these 
granules alternate with rows of cilia. He interpreted them as the 

^ Vitalfiirbung mit Neutralrot an Protozoen- Z. wiss. Zool. Bd. 63, 1898. 
^ Studien iiber Thigmotaxis bei Protisten. Arch. f. Anat. u. Phys. 1900. 

438 Gary N. Calkins. 

papilliform external swellings of the trichocysts and as merely 
condensed peripheral portions of the cortical plasm. My own 
observations support those of Wallengren. 

The third type of granule is interpreted by Prowazek as a 
ferment or enzyme bearer. Putter, on the other hand, believes 
them to be "respiratory granules" owing their staining capac- 
ity to the contained carbon dioxid. Wallengren's observations 
on starving forms led him to the belief that neither interpretation 
is correct, for, he argued, these granules being the first to disappear 
in hungry forms must be of the nature of stored food (see Figs. 
23 and 24). 

The crystals which are found in well-fed forms were identified 
by Schewiakoff as metaphosphate of calcium. They are of 
various forms and sizes and are confined to the endoplasm; being 
crystalline in nature they cannot be mistaken. They are now 
generally regarded as late metabolic products resulting from pro- 
teid digestion. 

b. The Ectoplasm. As in the majority of holo- and hetero- 
trichida the ectoplasmic modifications are well diflPerentiated from 
the endoplasm. A cuticle and underlying cortical plasm may be 
made out, the latter consisting of a much more dense substance 
than the endoplasm, analogous, probably, to the ectoplasm of an 
amoeba. In it are embedded the characteristic trichocysts 
which ordinarily project ever so slightly from the surface, giving 
rise to the minute papillae which may be distinguished in profile 
between the furrows of the cilia (shown in Fig. 20). In Para- 
moecium taken fresh from the pond water, the fixing agent which 
I have used, preserves the trichocysts within the cortical plasm, 
but after a few months under cultivation these organs cannot be 
made out, and seem to have been discharged and lost under the 
stimulation of the fixing fluid. Wallengren believes that they are 
taken into the endoplasm and digested as food in starving forms, 
but in preparations made from my cultures they are absent in the 
well-fed forms as in the degenerate ones. In all cases the spaces 
that were occupied by the trichocysts are present in the cortical 
plasm as vacuoles, and it is in this state that the relation to the 
peripheral papillae can be easily made out (<:/. Figs. 13, 18, 20 

Studies on the Life History of Protozoa. 439 

and 21). The difficulty appears to be that the cortical plasm is 
incapable of holding the trichocyst threads after expulsion, 
for the threads may be easily seen as a cloud around the 
animal immediately after fixation, while the after-treatment always 
dislodges them in the cultivated forms, but not in the wild forms. 

c. The Macronucleus. The structure of the normal macro- 
nucleus of ParamcBctum aurelia was described by Hertwig in 1889 
and the nucleus of P. caudatum agrees so closely with it, that 
further details are hardly necessary. It is an elongate, ellip- 
soidal body, usually with a smooth contour and without breaks 
of any kind save the minute impression made by the micro- 
nucleus. It frequently lies in a vacuole which is caused by 
the action of the fixing fluids, for in life the macronucleus is in 
immediate contact with the endoplasm. The contraction is 
probably in the endoplasm away from the nucleus rather than a 
contraction of the latter. Often there is a depression in the 
macronucleus due to the pressure of the contractile vacuole, and 
food vacuoles may also press against it, as Wallengren suggests, 
and distort it out of the normal proportions. 

In its finer structure the macronucleus is granular with the 
irregular granules densely packed together, giving the appearance 
of a homogeneous mass. ■ 

d. The Micronucleus. The micronucleus is usually embedded 
in the material of the larger nucleus, but may be separated from 
the latter, even in the resting stages, by a considerable distance, 
while in the dividing stages it is usually separated. Its finer 
structure consists of a more or less homogeneous mass of chro- 
matin frequently arranged in lines, while at one end is an accu- 
mulation of "achromatic" material in regard to which there is 
some diff"erence of opinion. In size the micronucleus is about 
1 1 microns, but in the difi^erent phases the size diff"ers so that this 
characteristic has but little weight. 

e. The Contractile Vacuoles. In the normal individual these 
are situated in the anterior and the posterior parts of the body, and 
about one-third of the length of the body from the ends. They 
are fed by radiating canals which are conspicuous in the living 
animals. The pulsation is regular as a rule, but this becomes 

440 Gary N. Calkins. 

spasmodic after prolonged captivity under a cover glass, and the 
irregularity is an index of the ultimate disintegration. 

2. Structure of Paronicecium m Depression Periods. 

a. Starvation and its Effects. The periodic depressions which 
were noted in the experiments, and which appeared at more or 
less regular intervals (viz: about every six months) were note- 
worthy because not always accompanied by the same type of 
degeneration as that characteristic of starved forms. 

A most comprehensive study of the structures of starved Para- 
mcecium was made by Wallengren, while various observers have 
called attention to the characteristic vacuolization which the cell 
protoplasm undergoes during starvation or at degeneration 
periods in any culture. In general, Wallengren found that the 
animals first use up the food material that is stored in granular 
form, in the endoplasm, and that when this reserve is used, the 
animals in lieu of other food, burn up first their endoplasm and 
then the cortical plasm. There results from this destruction, 
great vacuoles in the cell body which increase in size until the 
entire organism is distorted through the pressure of one confluent, 
or two or three great vesicles. Wallengren obtained his material 
by transferring the Paramoecia to tap water again and again, and 
thus ridding the medium of the customary bacterial food in a very 
short time. My own experiments to this end consisted in leaving 
the ciliates in a culture glass such as I have used throughout my 
experiments, until all the bacteria had been eaten and the culture 
medium had cleared. Thus a hundred or more individuals would 
be left for a period of a month or six weeks in the culture chambers 
where evaporation was prevented, and here they were watched 
daily until they ultimately died of starvation. While Wallen- 
gren's experiments were undertaken for the purpose of deter- 
mining the efi^ect of starvation upon all of the protoplasmic 
structures of these forms, mine were done for the purpose of 
studying the effects of such treatment upon the nucleus and endo- 
plasm, and general vitality. Wallengren found the following 
effects in the protoplasm of Paramcecium after starvation for a 
period of from 8 to lo days, which he designates as the "first 

Studies on the Life History of Protozoa. 441 

period" in the inanition phenomena: "All gastric vacuoles and 
food balls disappear during the first period. After this the small 
endoplasmic granules are used. As a result of this the quantity 
of endoplasm becomes much reduced. Toward the end of this 
time the living substance of the endoplasm itself is used, in part 
at least, to supply fuel for the continuing metabolism. Owing 
to the disappearance of the inclusions of the endoplasm and to the 
use of endoplasmic substance itself, the body form becomes more 
or less distorted or changed. But even in those individuals in 
which this has taken place and in which the form is considerably 
changed, the ectoplasm with its trichocysts, the contractile vacuole 
and the cilia are not altered in any noticeable manner. This 
shows, therefore, that in the first period of inanition the first 
materials to be used are the reserve stuffs which are normally 
utilized for the ordinary fuel (life processes). Only when all of 
the reserve material is used and when the endoplasm itself is first 
attacked, and only when all food whatsoever is gone, will other 
parts of the protoplasmic structures be attacked. When this time 
comes the second period is inaugurated." (Loc. cit., p. 87.) 

In the second period of inanition there are more fundamental 
changes, and the remainder of the protoplasmic structures are 
involved. Ultimately the nucleus is affected and when this goes 
the organisms are doomed. Wallengren's conclusions as to this 
period are as follows: "The endoplasm, which at the beginning of 
this last period was already considerably reduced, now appears 
strongly vacuolized. The shining vacuoles which are probably 
filled with the products of degeneration of the endoplasmic con- 
tents, may attain to a considerable size. Along with this vacuol- 
ization the ectoplasm becomes more and more absorbed and as a 
result of this, the trichocysts are drawn into the endoplasm 
streams and are probably digested. Along wnth them the small 
papilliform swellings on the outer surface disappear, and with 
these the small granules which in the living animal stain with 
neutral red. The contractile vacuoles and their feeding canals 
become reduced in the same proportion as the thinning of the 
ectoplasm. A number of cilia are probably absorbed as a result 
of the decreasing size of the whole animal, and the remainder of 

442 Gary N. Calkins. 

them are shorter than the normal. Owing to the inner changes 
the whole organism may at this time be so modified that it is 

"Thus during the continued inanition of the body, first one part 
and then another becomes absorbed, first the endoplasm, next the 
ectoplasm, the trichocysts and the cilia in part, all to maintain as 
long as is possible the vital functions. In the meantime, however, 
the nucleus has not escaped without changes as follows:" (loc. 
cit., p. 98) . . . "In the inside of the macronucleus a 
rounded mulberry-like mass is developed. Its alveolar structure 
has changed at the same time, and in the center there are usually 
one or two small central bodies (Binnenkorper). The high 
pressure which is developed in the decreasing body form and due 
to the enlarging vacuoles, causes the nucleus to become greatly 
deformed and compressed. The various parts of the nucleus are 
broken up into fragments which may probably be used more or 
less as food ( t). Of the former large macronucleus there is now 
left unchanged only the nuclear body which has been formed and 
this lies between the broken down nuclear parts." (Id., p. 112.) 
"In the micronucleus no destructive changes are mani- 
fested during the hunger degeneration. It is the one part of the 
body which is apparently not affected by the conditions of the 
experiments, a not unnatural result considering the importance 
which this organ of these cells has in rebuilding the macronucleus 
after conjugation. Of all organoids the micronucleus would thus 
seem to be the most important of the cell." (Id., p. 114.) 

These careful observations and clear results of Wallengren, 
most of which I have been able to verify, offer a good basis for the 
comparison of structures obtained in the different stages of the 
life history of Paramcectum {cf. Figs. 22 and 23). We may 
distinguish two types of degeneration changes in the series from 
the start to the finish. One set accompanies starvation, and was 
characteristic of the first two periods of depression, the other 
accompanies physiological depression of a different type at the 
last two periods. In the former the changes in structure had to 
do mainly with vacuolization of the endoplasm and rupture of the 
macronucleus, while in the latter the endoplasmic portions were 

Studies on the Life History of Protozoa. 443 

degenerated in a different way. The ectoplasmic parts and the 
micronuclear structures were not affected until the last depression 

The first clearly marked period of depression came in July, about 
six months after the cultures were started. It was characterized 
by a well-defined reduction in size (down to 109 microns; see 
Fig, 3), and by vacuolization of the endoplasm while the ecto- 
plasm did not appear to be much involved. Many of the individ- 
uals were characterized by great vacuoles similar to those in 
starved forms, which distorted the body almost out of recognition, 
in others the nuclei were fragmented into two or three parts, and 
in all there was a marked absence of the larger food granules and 
gastric vacuoles which characterize the normal animals, and this, 
notwithstanding the fact that bacterial food was present in abund- 
ance (see Studies I). As stated in these Studies (III) the 
organisms under these conditions still take food and in some cases 
the endoplasm appears opaque with the undigested food balls, but 
the decrease in size continues and the endoplasmic vacuolization 
is not prevented by the presence of the food. It is the digestive 
function, apparently, which becomes ineffective at such periods, 
and if this is a correct assumption, this function can be stimulated, 
as I have shown by the experiments. 

Identical results were obtained in the period of depression in 
December, 1901, a depression which was again overcome by the 
use of beef extract, while the individuals of the series which had 
been continued on the hay diet, all died. These became smaller 
and smaller, and again gave morphological indications of starva- 
tion, notwithstanding the fact that the individuals which had been 
stimulated with the beef extract were living and reproducing 
normally in the same food medium. They became much reduced 
in size, the endoplasm became distorted with vacuoles, and they 
died with absolutely no indication of disease through parasites. 

These observations show, therefore, that starvation effects may 
be produced even though the animals are living in a medium rich 
in food. It is trite to say that to prevent starvation we must have 
not only food but the ability to digest and assimilate it, yet com- 
mon as this observation is, it is important in the present connec- 

444 Gary N. Calkins. 

tion and involves a factor which cannot be overlooked in any dis- 
cussion on old age. 

In the June period, as stated previously, the same conditions 
w^ere not observed, for the organisms, in part at least, had been 
treated with the beef extract every week during the first three 
months, since the previous period of depression. The division 
rate began to run down in the case of the B series in April, in the 
A series in May, and in all of thq material that had been continued 
on the beef, the characteristic structure was a densely granular 
endoplasm (Fig. 7). In the specimens that had not been treated 
with the beef since the preceding December, this character of the 
endoplasm was not noted. These unstimulated individuals died 
out in about the 508th generation (B series) after becoming much 
emaciated and reduced in size, and with reduced nuclei. The 
nature of the protoplasmic changes is indicated, in one case at 
least, in Fig. 14. Here the macronucleus has entirely disap- 
peared, not even a granular trace remaining, while the endoplasm 
is crowded with vacuoles of considerable size. The micronucleus 
is slightly hypertrophied and has a very peculiar outer membrane 
within which the chromatin and achromatic material lie in what 
appears to be the real nuclear membrane. The dense granules 
characteristic of the beef-fed individuals are absent. The un- 
stimulated A series did not die out until about two weeks later. 
At the time when the B individual described above died (May 12) 
the unstimulated A series was characterized by somewhat reduced 
size, a declining division rate, and absence of the dense protoplas- 
mic granules. In the stimulated A series, on the other hand, 
(Ai and A2) of about the 560th generation, the structures were 
normal, gastric vacuoles were numerous and divisions were fre- 
quent. Towards the end of June, however, when the A series 
nearly died out in the 620th generation, the conditions were very 
different. Fig. 7 is from a specimen in the 615th generation. Its 
size is below the normal; its endoplasm is choked up with granules 
and there is no trace of vacuoles save the contractile vacuole near 
one end. The macronucleus is definitely granular, and its con- 
tour 'is irregular as though devoid of nuclear membrane. The 
micronucleus is elongate and spindle-formed. The ectoplasm is 

Studies on the Life History of Protozoa. 445 

not deformed and save for the absence of trichocysts it appears to 
be normal. This was the condition of the protoplasm when the 
usual large number of culture individuals was reduced to 6 A's 
and no B's, and a condition from which the A series were rescued 
only with the greatest difficulty by the use of pancreas extract. 
Figs. 8, 9 and 10 represent individuals that had been in extract of 
pancreas for 48 hours, and then transferred to hay infusion. They 
are identical, therefore, with the individuals that lived and carried 
the race to the 742d generation. In these forms the endoplasm 
in most cases is normally vesicular in the center and gastric 
vacuoles are common, while the ends alone still retain the dense 
granular aspect. 

From this time until the race died out the division rate was 
sluggish. The conditions of the protoplasm in the later individ- 
uals was decidedly characteristic (Figs. 17, 19, 20, 21 and 22). 
Throughout the fall, individuals would appear with densely granu- 
lar protoplasm, which is invariably the sign of death, unless the 
animals are stimulated in some way. In such forms the macro- 
nucleus may or may not be normal, whereas the micronucleus as 
a rule becomes hypertrophied and the ectoplasm full of great 
vacuoles. Fig. 22 is a good representation of the conditions 
at this time. The endoplasm is apparently normal; there are 
food vacuoles and endoplasmic granules, and vesicular structure, 
but the micronucleus is spherical and vesicular, has lost its usual 
place in a niche in the macronucleus and shows evidence 
of granular modification of the previously homogeneous chro- 

The sister-cell of the one pictured in Fig. 22, and one of the two 
oldest of the A series (742 generations), showed the following 
points while alive: "A12 was alive this morning and was picked 
out for examination. It had two contractile vacuoles situated 
dorsally and close together. The astral canals were absent; in 
their place was a row of dorsal feeding canals, such as those 
characteristic of the more generalized holotrichida {e.g., Chlamy- 
dodontida:). The rest of the body contained eight or ten large 
vacuoles not contractile. The macronucleus was slightly hyper- 
trophied, and visible, indicating the approach of disintegration. 

44^ Gary N. Calkins. 

The papillae of the cuticle were plainly visible and what T have 
taken to be apertures of the trichocysts were more or less 
numerous. (This is shown in the preserved sister-cell, Fig. 22.) 
A few trichocysts remained in the cortical plasm, but there were 
many vacuoles in this layer indicating that when the trichocysts 
were discharged they were not re-formed. The peristome was 
normal and the mouth had a vigorous oral membrane. The 
size was large, fully as great as any of the preparations that 
had been made at any time during the 742 generations. Move- 
ments vigorous to slow, with a tendency on the part of the animal 
to remain stationary." ^ 

It was while the organisms were in this structural condition that 
the many attempts to rejuvenate the race were made as described 
in the previous pages, and it was in this condition of the proto- 
plasm that the race finally died out from exhaustion. Before 
dying, however, the individuals, as indicated in the above para- 
graph from my notes, were of full size and were filled with gastric 
vacuoles and partly digested food, while the body form was 
normal, (compare Figs. 2 and 21). 

It must be admitted that these forms were capable of individual 
growth at this period and, since the macronucleus was normal in 
the last individuals luhile the micronucleus was considerably 
changed, it must be further admitted that the vegetative metabolic 
processes were presumably reinvigorated; on the other hand, the 
functions of reproduction; that is, of division, were degenerated 
possibly, if not probably, because of the apparent degeneration of 
the micron-ucleus and of the cortical plasm, whose functions were 
not reinvigorated by the artificial means which were tried. 


Although only a beginning has been made to determine the 
objects for which this series of experiments was started, it is 
advisable to bring together the results thus far attained and to see 
how they conform with the a prion conceptions which were cur- 
rent at the outset of the experiments. 

1 From my note book. 

Studies on the Life History of Protozoa. 447 

It is not out of place to consider first the initial objects of the 
undertaking, although at the risk of again repeating what has been 
often stated. 

1. The first aim of the experiments was to get light upon the 
general phenomenon of conjugation and through this, upon fertil- 
ization in general. 

2. To determine whether conjugation is imperatively necessary 
for rejuvenescence. 

3. To determine whether artificial rejuvenescence is possible. 

4. To determine the conditions, antecedent and subsequent to 

5. To determine, if possible, the significance of rejuvenescence. 

6. To determine, finally, whether protoplasm in these simple 
forms is capable of indefinitely continued life without conjugation, 
or whether it is subject to the conditions of "old age." 

On none of these points can a definitely positive answer be 
given, and further experiments must be undertaken to clear them 
up. The fact that, after a continuous cultivation of 742 genera- 
tions, covering a period of 23 months, the race died out apparently 
from exhaustion, shows that under the conditions, continued life 
was impossible, and if this conclusion, which seems to be the only 
one justified by the results, be granted to obtain in nature, then 
we must agree with Maupas and others that the indefinite con- 
tinuance of life without conjugation, is improbable. 

I. The Conditions of the Experiments. 

The question has been raised whether the conditions under 
which the experiments were undertaken were in any way abnor- 
mal to Paranicecium, and whether, from the results obtained, we 
are justified in carrying the conclusions to the free-living forms, 
and to similar types in general. 

It might be objected that the space allowed was inadequate; 
or, that the light conditions were abnormal; or, that the water 
would get foul; or, that they were given only one kind of food; 
or, that they were subjected to pressure. If we examine these 
objections critically we shall find that they have little basis. 

Let us consider first the matter of space, for this involves some 
of the other objections, viz: pressure, volume conditions, and the 

448 Gary N. Calkins. 

like. The actual amount of water that was used for each isolated 
individual was one-half a cubic centimeter. This was contained 
in a small chamber consisting of a hollow-ground slide, two glass 
supports about 3 mm. thick, and a thin glass cover. The Para- 
incecium had ample room, therefore, for free movement, and an 
actual depth of water of more than an eighth of an inch. Pres- 
sure, therefore, was out of the question. In such a slide chamber 
individuals were kept (/. e.^ extra individuals from the "stock") 
for periods considerably longer than six weeks without change of 
water, showing that the mere quantity was sufficient in order to 
keep the animals alive. Foulness of the water, accumulation of 
carbon dioxid, lack of oxygen, etc., were all guarded against 
by the almost daily transfer of the culture individuals into fresh 
hay infusion. The salt content of the water remained practically 
constant, for fresh hay infusion was used each time with the same 
amount of hay from the same source while the weekly analysis 
of Croton water shows only minor fluctuations in the small 
quantity of salts in solution. The gradual decrease in vitality 
cannot be attributed to these causes, a similar phenomenon being 
a matter of common observation and noticeable in any culture of 
protozoa, no matter how large the vessel, nor what the species. 
The light conditions were similar to those in any laboratory, the 
culture vessels being kept before a window with north exposure. 

In regard to the possible objection that the Paramcecium ob- 
tained only one kind of food, and therefore that the conditions 
were abnormal in this respect, it may be stated that such a condi- 
tion of treatment is a sine qua non of the experiments, and the only 
possible means of controlling the results, and as I have demon- 
strated, it is by a change of diet, including salt constituents, that 
the periods of depression are overcome. This objection, there- 
fore, begs the question of an object of the investigation. 

It seems quite unnecessary to repeat again that the only normal 
life possible to Paramcecium caudatum is in the ponds where it is 
subject to the changes in chemical composition of the water, to 
the exigencies of drought, of heat, of freezing, and of rest by 
encystment or lack of food. In the laboratory the protoplasmic 
activities get no rest, but day after day they are maintained at the 

Studies on the Life History of Protozoa. 449 

optimum rate and such conditions can by no stretch of the imagin- 
ation be called identical with those of the ponds. Yet the "nor- 
mal conditions" may, after all, be but a matter of definition. If 
we leave a hay infusion to stand exposed to the air, Paramcecium 
will ultimately appear in it, and will ultimately die out from it. 
The appearance and disappearance cannot be called artificial, it is 
as much normal for Paramcecium to appear in such an infusion as 
it is normal for the bacteria upon which the animals feed to be 
there. City life for man may be called an artificial life as opposed 
to the "normal" original or pastoral life, but it is no less normal 
now than the primitive life was, even if it is found that the average 
length of life is shorter for urban than for country-dwelling people. 
The course of human life, or the history of the race, physiologi- 
cally speaking, is no less normal for being rapid. In the same 
way we may argue for the race of Paramcecium and its life in the 
culture chambers of the laboratory. The life pursues a normal 
course, although possibly faster than in nature, and the ultimate 
results obtained in cultures may be confidently expected to obtain 
sooner or later in the natural habitats. Seven hundred and forty- 
two generations represent a long time for organisms to live and 
develop in a medium that is not normal, and the mere fact that 
they do so live is sufliicient evidence to prove the point. It seems 
to me, therefore, perfectly legitimate to take the phenomena of 
vitality in Paramcecium in culture as practically identical in 
outcome with the phenomena in natural surroundings, and as 
indicative of what goes on in living protoplasm under "normal 

Looked at from this point of view, the experiments teach (i) 
that a given form together with the race derived from it will 
exhibit periodic depressions in vital activity; (2) That such 
depressions can be overcome by artificial means ( and probably 
but not surely, in nature by opportune changes in the immediate 
environment). Further than this, however, the experiments 
teach, (3) that these depressions are not all of the same type, nor 
due to the same causes. They give reason for the belief that 
periods of depression may ensue wherein different functions give 
out, and that when this occurs, as for example when the cortical 

45 o Gary N . Calkins. 

plasm and micronucleus show evidences of degeneration, all of the 
experiments that we may try to artificially reinvigorate them, 
will probably be futile. This may indicate one of two things, 
viz: that under natural conditions changes m immediate environ- 
ment would be insufficient to rejuvenate when the organisms are 
in this ultimate state of exhaustion, unless, indeed, the experi- 
ments failed to eliminate all of the chances such an organism bas- 
in nature; or, conjugation is a necessary condition of continued 
protoplasmic activ.ty. 

I am inclined to the belief that some material might ultimately 
have been found which would have helped the Paramoecia over 
this period of extremity and would have stimulated micronucleus 
and cortical plasm to continued work. The failure to find it, 
however, indicates a like difficulty in nature and makes the 
a priori reason most probable that the phenomena with which we 
are familiar, namely, the processes of conjugation, have been 
essential in mamtaming the races of Paramecium up to the pres- 
ent time, and in keeping them from extinction. 

2. Does Protoplasm Grozu Old? 

The above considerations lead to the discussion of age in a 
simple cell organism. In higher forms old age is manifested by 
the gradual weakening of the vital functions, waste matters are 
inadequately disposed of, or are retained in one form or another 
in the cells and tissues; this involves the physical impairment of 
organs and enhances the difficulties of their functional activities 
until, by the accumulation of such mutually aggravating processes, 
the organism ultimately dies of "old age." In Paramcectum 
there is little morphological evidence of the onset of old age, 
although, if we accept the impairment of the vital functions as an 
index, we must conclude that diminution of the division rate, 
decrease in size, etc., are evidences of this phenomenon in pro- 
tozoa. So far as the accumulation of waste matters is concerned, 
there is morphological evidence to indicate that this takes place 
more frequently at periods of depression. There was no sign of the 
crystals which frequently accumulate in the protoplasm of various 
protozoa, and in the last specimens of the race (742d generation) 
both endoplasm and macronucleus were normal in structure. 

Studies on the Life History of Protozoa. 45 1 

The surest evidence of what may be considered old age 
in this form, was therefore, functional, and was expressed by 
diminished division rate and by the increased frequency of ab- 
normal binary fission. Abnormal division, as a matter of fact, 
like nuclear hypertrophy, may occur at all periods and marks 
some particular weakness of the single individual; occurring more 
frequently, however, at certain periods of depression, such ab- 
normalities give evidence of general protoplasmic weakness. 
The various types of incomplete division are very instructive and 
a more prolonged study than I was able to give to them might 
afford positive evidence of the nature of the pathological changes 
involved. In some of the specimens which T obtained during 
periods of depression, the macronuclei and micronuclei appear 
normal; in others there is a macronucleus in each of the daughter 
individuals, but the micronucleus is undivided; in others the 
macronucleus is divided but remains in one individual, the micro- 
nucleus is undivided and remains with the original macronucleus, 
while the daughter individual has no trace of nuclei. In all cases, 
finally, of pathological division the cortical plasm appears ab- 
normal and vacuolar, while the endoplasm is very frequently 
disintegrated and abnormal (Figs. 25 and 26). 

While these observations are too few to permit far-reaching 
conclusions, they are sufficient to indicate that some protoplasmic 
changes have taken place, and further, that the cortical plasm has 
become modified in some way. Indeed, the inability completely 
to divide may be accounted for by the loss of vitality in this par- 
ticular part of the protoplasm, for in the majority of cases the 
initial stages of division are safely passed, the final separation 
alone being retarded and usually omitted altogether, so that 
monsters of three or four individuals may be formed through the 
continued incomplete division of the original degenerate specimen 
(Fig. 25). As is well known, the cortical plasm is the seat of the 
myoneme formation, of the cilia, and of other motile organs, and, 
in general, may be said to possess kinetic or motor functions. 
That this portion of the protoplasm is subject to change is shown 
by the fact that at certain times the outer protoplasm becomes 

452 Gary N. Calkins. 

sticky or plastic and to such an extent that two individuals upon 
meeting, fuse together in plastogamy. This, which I have 
termed elsewhere the "miscible state," may be so marked that 
groups consisting often of from five to eight aggregated individuals 
are occasionally seen. It is analogous, apparently, to the plas- 
togamy so often seen in the fresh water testacea such as Diffliigia 
or Arcella, which Schaudinn has recently shown to have no 
connection with conjugation in these instances. In Paramcecium 
during this miscible state, conjugations are for the only times 
possible, and many complete conjugations are found together 
with the fused multiple individuals. There is no doubt, then, that 
the cortical plasm changes in physical condition, and there is equal 
reason to believe that at periods of depression when these abnormal 
divisions are more frequent, the cortical plasm shows degenerate 
conditions^ or, possibly, a condition of old age. 

There is therefore some significance in the fact that the cortical 
plasm gives out; some significance connected with the diminishing 
division rate and with advancing old age as evidenced by dimin- 
ishing activity. 

While there may be some uncertainty as to whether the decreas- 
ing vitality of a race of Paramcecium is evidence of normally 
decreasing functions indicative of protoplasmic old age, or of 
some other cause of degeneration, there is absolutely no reason to 
believe that it is due to a parasite of any kind, nor to any harmful 
substances in the medium in which they live. In the earlier 
periods of depression there seems to have been a gradual loss of 
powers connected with metabolism, and of something which was 
vitally important to the race, for unless the individuals were stimu- 
lated, they inevitably died. This was strikingly demonstrated in 
the period of depression in December, 1901, when a number of 
individuals of the regular series were continued on the usual hay 
infusion, while others were treated with beef extract for 24 hours, 
and still others with salts of different kinds for not more than 30 
minutes. The non-stimulated forms showed increasing sluggish- 
ness and depression, and all died in the course of two weeks, while 
the sister-cells which had been stimulated, lived with varying 
fortunes until a year from then (see Diagram I). The pertinent 

Studies on the Life History of Protozoa. 453 

questions may be asked was it old age from which the organism 
died ? and, if so, what form did it take ? They were fed daily with 
the same food upon which the stimulated sister cells thrived, but 
they could not assimilate it and would not grow nor divide. In 
similar cultures which had been carried to a like point by previous 
observers, the entire race died, and although no evidence of 
structural degeneration was evident, it has been taken for granted 
that their organisms died from exhausted vitality, or in other 
words, of old age. In my cultures there was some evidence of 
degeneration, especially in the endoplasmic structures and in the 

The fact that stimulation was successful in carrying the race 
through this earlier period of depression indicates either that the 
conditions are not the same as those accompanying old age in 
metazoa, or else that such conditions may be satisfactorily over- 
come. I believe the conditions are more or less the same in both 
cases, and that in senile Paramcecta certain functions have become 
retarded, possibly by the accumulation of useless protoplasmic 
elements too minute to be detected, or by some less mechanical 
cause connected with the molecular structure of protoplasm and 
which, therefore, affords no morphological evidence of change. 
Such an hypothesis would explain the difference in length of time 
required to get positive results in the stimulation experiments. 
For example, in August, 1901, after the race had been on hay 
infusion continuously for 7 months, it was necessary to keep the 
single individuals on beef extract for three weeks before they 
would live again in the hay infusion. But in December it was 
necessary to keep them on the stimulant only a day or two to get 
the desired result. The short treatment at this period sufficed, 
because they were not allowed to become weakened to the same 
extent as in the preceding period of depression. This result 
points to some physical condition of the protoplasm, possibly to 
the accumulation of some protoplasmic product or products which 
lead to diminished vigor and to death. Reinvigoration after a 
more or less prolonged treatment with the beef extract and stimu- 
lation by this and other means indicates that such materials are 
disposed of, or, more generally speaking, and to use a phrase which 

454 Gary N. Calkins. 

Wilson in The Cell attributes to O. Hertwig, that the condition of 
stabiHty is changed into one of protoplasmic lability. 

While such an hypothesis accounts for the first two periods of 
depression, it fails to account for that of June and of December, 
1902. In the interval between June and the preceding December, 
the race in part, had been treated weekly with beef extract until 
the first of April, after which the organisms had been fed with the 
usual hay infusion. In June they began to degenerate, and from 
this time on, treatment with the beef extract was futile, and the 
race was finally saved only by using extract of pancreas and of 
brain. This, however, gave only temporary relief and complete 
activity was never again recovered and the division rate remained 
below the average, until the race finally became extinct in Decem- 
ber, 1902, and this despite the fact that, morphologically, the 
endoplasm and macronucleus were restored. 

Was the last period of depression running from June until 
December, 1902, an expression of old age ? From the structures 
of the organism and their behavior, there is no doubt that the 
ailment at this period was different from that of the earlier 
periods of depression, and there is no doubt again that the reme- 
dies which had succeeded at the earlier periods failed completely 
at this. The final depression of vital activities may be accounted 
for by one of two assumptions : (i) There was an accumulation of 
waste material of a different kind from that of the earlier periods, 
or a different physical condition, and a weakening of different 
functions, or (2) certain elements in the protoplasm endowed with 
a given potential of activity used up that potential and failed to 
recover it by artificial stimulation. Or a third hypothesis may be 
conceived which embodies both of these. The morphological 
structure at the final period shows that some different elements of 
the body were involved in the last period of depression, and that 
the elements which had given out in the previous periods were 
satisfactorily reinvigorated even in the last individuals of the 
race. Thus the micronucleus and the cortical plasm showed 
unmistakable signs of degeneration in the last few individuals of 
the race, while the endoplasm and macronucleus were perfectly 
normal in appearance, and metabolism, which these elements of 

Studies on the Life History of Protozoa. 455 

the cell appear to control, seemed to be equally normal, since the 
organisms were of full size, while the endoplasm was full of 
partly digested food. It appears, then, that the experiments 
were successful in reinvigorating the elements of the cell that had 
given out in previous periods of depression, but that other ele- 
ments were now involved which all my experiments failed to 
reach. Here a more deeply-lying malady had to be met, and the 
experiments being unsuccessful in meeting it, the entire race 
died out. This series of facts appears to warrant the assump- 
tion that there is a fundamental difference in the protoplasmic 
elements which go to make up the body of a protozoan, one of which 
IS to be compared with the somatic cells of metazoa, the other tvith 
germ cells; the one connected with the vegetative functions of 
metabolism, the other with reproduction; the one may give out and 
so lead to "physiological death" (Hertwig) or it may be restimu- 
lated; the other may give out and so lead to "germinal death" of 
the race. 

It is not outside the range of possibility that the last depression 
period might have been overcome by some suitable experiments, 
and the fact that we did not succeed in finding a suitable stimulant 
does not justify us in assuming that this period represents the last 
vital spark of this protoplasm, any more than we are justified in 
assuming that the earlier periods of depression represented this 
condition. If, however, some element or elements of the proto- 
plasm become exhausted and all experiments to replace them fail, 
then we might justly speak of exhaustion or "old age" of these 
elements of the protoplasm and affirm that old age in one form, 
characterized the organisms during the first two periods of de- 
pression, while it took another form in the final period. 

3. Conjugation and Rejuvenescence. 

"Old age," then, appears to be a natural condition of living 
protoplasm and we may ask, is there any experimental evidence 
to show that this condition may be overcome by natural means ? 

It has been generally assumed by biologists that conjugation 
brings about rejuvenescence in the conjugating individuals, and 
so imparts to the ex-conjugants and to their immediate descend- 
ants a high potential of vigor. During the process of conjugation 

456 Gary N. Calkins. 

there is a complete change of materials or of protoplasmic make- 
up, and a thorough "reorganization," to use the excellent term 
proposed by Engelmann. Not only is there a new conjoint micro- 
nucleus with its chemical compounds derived from the union of 
two nuclei from individuals of diverse environment, but the endo- 
plasm and cortical plasm must receive new materials through the 
disintegration and the absorption of the old macronucleus, and of 
at least three-quarters of the old micronucleus. If, as I have long 
maintained, there is a specific "kinetic" substance in the proto- 
zoon nucleus, a substance which in the centro-nuclei forms the 
division-center and which is found in the micronucleus of Para- 
moecium, then the cytoplasm of Paramcecium must receive a 
certain amount of "kinoplasm" at each period of conjugation 
and from the experiments, enough to carry the race through a 
complete cycle. In my cultures such reorganization by conjuga- 
tion was prevented in the straight line of the experiments, and the 
only opportunity for reorganization came with the change in diet. 
This, indeed, seemed to be operative for some time, but ultimately 
failed, as we have seen. In the stock material, however, material 
left over after the individuals had been selected for the cultures, 
conjugation experiments were frequently tried during the course 
of the experiments, and the results have been given (Studies I). 
Some of the results are very suggestive in the present connection, 
for it was found that only a few of the ex-conjugants continued to 
live, approximately 6 per cent of them. This result may be due, as 
I have previously stated, to the fact that both of the gametes had 
been under identical conditions of food, etc., and no new sub- 
stances were formed by the union of similar nuclei and protoplasm. 
Or the result may be due to the fact which Stevens^ calls attention 
to, that conjugation is an exhausting process, and that, being 
weakened through long cultivation in cultures, these ex-conjugants 
did not have sufficient vitality to recover. This suggestion does 
not set aside all of the difficulties, however, for we have still to ex- 
plain the large number of cases where the ex-conjugants have lived 

^N. M. Stevens Further Studies on the Ciliate Infusoria, Lichnophora and 
Boveria. Arch. f. Protistenk., Ill, 1903. 

Studies on the Life History of Protozoa. 


DEC. JAN. 1 rea. 

1 1902 1 1 1 


1 1 

1 ' 1 1 

1 1 


1 , 


1 ' 


1 1 


1 , 





1 ' 







r— 1 


1 1 



1 1 





' 1 

' 1 1 1 1 


' [ 1 1 1 

■ 1 ' 1 j 1 


1 ' 1 1 1 


1 1 1 1 1 

1 1 ■ 1 ! 

Diagram II. 

Complete history of the endogamous ex-conjugant by ten-day periods, 
and periods the same as in Diagram I 



1 1 1 1902 1 1 1 

} JULY 1 AUG. 1 SIPT. 1 OCT. | NOV. I DEC. | JAM. 


90S • 1 " 
MAR. j APR. 1 MAY 1 

n r 




' ' 


t 1 ! • ' 



! ' : 

' ' ' 



'Hi: I 

1 L_r- 

! 1 ■ 1 




! ' ! 1 


1 r 


1 1 

' 1 



' 1 1 1 ' 


, , 




1 ' ' 



' 1 ' 1 I ' 1 





1 ' 1 1 1- ' ' 


i ' 1 1 ' ' • 


1 ' ' 1 I ! ! 

1— , 




1 1 1 j ■ 




1 . 1 1 1 1 1 




I ! • ' 




! 1 

Diagram III. 
Complete history of the third series (C) by ten-day periods Ordinates and 
periods the same as in Diagram I, As in Diagram I, this curve- represents the 
average division rates of four lines of individuals 

458 Gary N. Calkins. 

and reproduced for from 8 to 20 generations, and with apparently 
well-organized bodies/ 

One of these successful cases was an ex-conjugant from an 
endogamous union of two individuals which were separated by 
not more than eight or ten generations from the ancestral A in the 
354th generation of my cultures. The other ex-conjugant died 
out in the iith generation, while the successful one ran through 
376 generations before showing signs of debility. It went through 
eight months in culture without beitig stimulated, and died out 
finally at the end of 376 generations, which was exactly three 
generations less than the life of the third series of Paramcecium 
(C series) which I started on June 18, 1902, and carried along in 
culture until May 30, 1903, when it died out in the 379th generation 
(see Diagram III).^ 

Unfortunately, this ex-conjugant has not an absolutely clear 
record, for the first day after the pair had separated, I placed them 
both in beef extract for 24 hours (December 9, 1901). This 
experiment had failed a number of times, and I had no reason to 
believe that it would succeed this time, and, as stated above, one 
of the two ex-conjugants thus treated died after eleven generations. 
Although at first I attributed the successful result to this treat- 
ment, I do not now believe that the beef extract had anything 
to do with the vigor of the race that followed, and believe 
that rejuvenescence was accomplished by the conjugation and 
nothing else. This conclusion is based upon the following facts: 
(i) Other ex-conjugants similarly treated with the beef extract 
failed to live; (2) the non-con jugating individuals of the regular 
series which were treated with the beef extract at the same time 

'See Studies I, table of conjugations opposite p. 174. 

^We have, then, the interesting coincidence of an individual running through 354 
generations in culture, conjugating with one of its OAvn close relations, and then, as 
an endogamous ex-conjugant, running through 376 generations more, a total of 730 
generations. Against this we must set the 742 generations of the main culture 
series, and the 379 generations of the third series (C) . The close connection between 
the 379 and 376 is very significant, and were it not for the fact that the first two 
series, A and B, were at a fatal period of depression at the end of 200 and 190 genera- 
tions, we might conclude that 370 more or less is the normal length of life of Para- 
mcecium in culture. 

Studies on the Life History of Protozoa. 459 

had a lower division rate and died out before May 5; that is, 
after running five months {cf. Diagrams I and Iiy It follows, 
therefore, that something luas operative in the ex-conjugant that 
was absent in the stimulated form, and this something could 
be nothing else than the reorganization which follows conjuga- 
tion. The accompanying curves show that the periods of 
depression and death which menaced the regular series in 
December, 1901, and again in June, 1902, were not paralleled 
in the cultures of the descendants of the ex-conjugant, and 
the conclusion is obvious that conjugation provided some stimu- 
lus which enabled this line of Paramcecium to live through 
periods in which the allied races were saved only by vigorous 
treatment and stimulation. There is no doubt at all that, had 
I tried to revive the race of the ex-conjugant by beef extract 
at: the end of August, 1902, I could have done so, for there 
was nothing serious in the nature of the depression at this time, 
when I allowed them to die without making an effort to save 
the race. It is now a matter of deep regret to me that I did 
not try to save them, and see if they would live beyond the time 
when the allied lines died out in December, 1902. Had they done 
so, it would have been still more convincing proof that conjuga- 
tion does actually rejuvenate and overcome the conditions of 
so-called "old age." I believe that the evidence which I have 
outlined above is quite sufficient, however, to establish this point, 
the one questionable factor being the beef extract, and even this, 
as I have shown, could have only a limited bearing and does not 
at all outweigh the positive evidence in favor of the conclusion. 

Columbia University, 
April, 1904. 

460 Gary N. Calkins. 


All of the photographs were taken by Dr. Edward Learning from permanent preparations of Para- 
mcecium caudatum, stained with picro-carmine. All are equally magnified and the relative sizes 
represent absolute differences. 

Plate I. 

Figs. I and 2. Two normal specimens B series (107th generation and after three months of cul- 
ture in hay infusion. These do not differ from typical Paramoecium from the ponds, and have many 
endoplasmic vacuoles, alveolar protoplasm, and homogeneous nuclei. 

Fig. 3. A typical individual of the B series during the first period of depression. The ectoplasm 
is fully as clearly defined, and as thick as in the largest forms, indicating that this portion at least, has 
not suffered from degeneration, a result differing from that in starved forms. (Compare Figs. 23 
and 24). 

Fig. 4. An individual from the B series in the 306th generation, stimulated with beef extract in 
August, fed continuously with hay infusion for three months until killed. The endoplasm is filled with 
gaGtric vacuoles and with partly digested food, the dissociated or "labile" condition of the endoplasm 
shown here is characteristic of Paramoecium under normal conditions. 

Fig. 5. An individual from the A series during the third cycle (550th generation), and twenty-four 
hours after treatment with beef extract. The endoplasm is filled with gastric vacuoles, the macronu- 
cleus is normal, but the micronucleus has divided three times and a clump of six nuclei may be seen at 
the lower end. There is a tendency toward a denser structure of the endoplasm, especially at the 
two extremities, this being indicative of approaching physiological depression. 

Fig. 6. An individual from the A series in the 560th generation. Treated 48 hours before fixation 
with beef extract. Gastric vacuoles are abundant in the upper portion, but in the lower part the 
characteristic density which marks the climax of physiological depression is shown, /. e., an apparently 
general "loading" of the protoplasm with inert material. 

Fig. 7. An individual from the A series in the 615th generation killed at a time of general depres- 
sion. It shows the typical condensed appearance when the power of division is lost and leads to death 
after several days without division. 

Fig. 8. An individual from the A series in the 623d generation (June, 1902,) and 24 hours after 
successful stimulation with extract of pancreas. The condition shown in Fig, 7 has been successfully 
overcome, and activity renewed by this treatment. This and the two following figures show stages in 
the breaking up of this dense, endoplasmic mass. The macronucleus is divided while the ends alone of 
the animal still retain the densely granular character. 

Figs. 9 and 10. Two individuals from the A series 48 hours after successful stimulation with 
pancreas extract. The endoplasm is now in a "labile" condition, although the extremities are still 
dense. The individual shown in Fig. 10 is further advanced in recovery than that shown in Fig. 9, 
but both are sister cells of individuals that carried the race to the 742d generation. 

Plate II. 

Fig. II. Two individuals of the A series in the 604th generation, two weeks prior to the fatal 
depression of June, 1902. These were treated with a weak solution of dibasic potassic phosphate (see 
paper) for 30 minutes and then transferred to hay infusion and killed 24 hours afterwards. 

Fig. 12. Two individuals of the A series treated at the same time as the preceding, with a weak 
solution of magnesium chloride for 25 minutes. The protoplasmic structures are normal and the endo- 
plasm has the typical alveolar appearance. 

Studies on the Life History of Protozoa. 461 

Fig. 13. Another individual treated with potassium phosphate. The micronucleus is separated 
from the macronucleus. At the lower end the focus is sufBciently sharp to show the characteristic 
papilli-form structure of the trichocyst spaces in the ectoplasm. 

Fig. 14. One of the last individuals of the B series (504th generation) in which the macronucleus 
is entirely gone. The micronucleus is distinct, and has its chromatin massed near one pole. The place 
which held the macronucleus is marked by a large vacuole. There are no observations to indicate the 
fate of this macronucleus, the break at the left side indicates that it may have dropped out at some period, 
although this did not happen during the course of the treatment, because the same condition was 
observed during its life, and immediately after killing. 

Fig. 15. An individual from the B series in the 502d generation after treatment with beef extract. 
The characteristic dense endoplasm is still present but there are many gastric vacuoles, while the micro- 
nucleus has divided three or more times and the daughter nuclei have accumulated at one end. 

Fig. 16. An individual of the A series in the 6o2d generation treated for 25 minutes with phos- 
phoric acid. It was transferred to hay infusion and killed 24 hours afterwards. The macronucleus is 
broken into fragments; the micronucleus has divided and one part (left center) seems to be forming a 
new macronucleus. (This individual offers the only evidence obtained of nuclear fragmentation and 
reconstruction through artificial means.) 

Fig. 17. An individual from the A series in the 718th generation, killed October 20 after living six 
days without division. The endoplasm shows a general absence of the larger granules, indicating 
starvation; the micronucleus (dimly visible at the lower end of the macronucleus) is hyaline and without 
chromatin, evidently degenerated. 

Fig. 18. Two individuals of the A series in the 6o2d generation, treated with a dilute solution of 
sodium chloride (see description), for 25 minutes. Transferred to hay infusion 24 hours afterward, 
and killed. 

Plate III. 

Figs. 19, 20, 21 and 22. Four individuals at the end period of the A series. Note that in all of 
these the macronucleus is not abnormally large as compared with normal individuals. This was true 
throughout the entire race at this period and contradicts Hertwig's recent theory of the size-relations at 
periods of depression. Fig. 19 represents an individual in the 720th generation, unusually small and 
unlike the remainder of the culture at this time. Fig. 20 represents an individual in the 725th 
generation, with conspicuously dense endoplasm and macronucleus. The latter bulges out towards the 
observer and the effect of the ectoplasm about it is that of a special nuclear capsule. Fig. 21, an 
individual in the 741st generation showing the looser texture of the endoplasm, gastric vacuoles and other 
characters, indicating that these organs had been restored by stimulation. The micronucleus is hyper- 
trophied, the macronucleus is normal. Fig. 22 represents an individual in the 742d generation, the 
oldest of the race. It shows the reorganized endoplasm, gastric vacuoles, and the like, but ectoplasm 
and micronucleus are degenerated. The former by vacuolization (note punctate appearance on right 
of macronucleus) the latter by hypertrophy and loss of chromatin. 

Figs. 23 and 24. These represent individuals which were starved for two and four weeks respect- 
ively. Those in Fig. 23 were fed on beef extract August 19, transferred to hay August 20th and left 
unchanged until September 19, when they were killed. The individual shown in Fig. 24 was not given 
beef extract, but was left in hay infusion for two weeks unchanged, when it was killed. In Fig. 23 
the spots at the lower ends represent the micronuclei, in Fig. 24 the upper elongated granule is the 

Fig. 25. A triple monster from an individual 72 hours after conjugation, with many nuclear frag- 
ments and evidence of two incomplete divisions. 

Fig. 26. A double monster from the A series in September, 1901. The micronucleus is undivided, 
the macronucleus is deeply cleft and the individual on the right has no trace of nuclei. 



The Journal of Experimental Zoology 



The Journal of Experimental Zoology 



The Journdl of Experimental Zoology 






With 47 Text Figures. 


It is beyond the purpose of the present paper to review the whole 
question of the relation between the nervous system and morpho- 
genesis or the sustaining effect of "trophic" stimuli upon form. 
It is hoped, however, that the observations and experiments to be 
described, together with the interpretation offered, may serve 
to throw some light upon this most interesting, but difficult 

The question of the relation of the central nervous system to 
regeneration in the lower animals has been touched upon by vari- 
ous authors. As regards Planaria, a species which has been the 
object of study by many investigators, opinions differ to some 
extent at the present time. It is well known that in this form 
removal of the cephalic ganglia does not interfere with complete 
regeneration, the ganglia themselves being regenerated from por- 
tions of the nervous system which may be present. As regards 
other parts of the central nervous system, however, Bardeen ('03) 
holds to the opinion that some portion of the nerve cords or of one 
of them must be present in order that regeneration may occur, 
while Morgan ('98, '00, '01, p. 44,) believes that regeneration may 
occur in pieces from the lateral region of the body which contain 
no part of the longitudinal cords. In an interesting paper dealing 

464 C. M. Child. 

with Phagocata and Dendroccelum, Lillie ('01) records the fact 
that while Phagocata equals Planaria in its regenerative power, 
the conditions in Dendroccelum are widely different. In this 
form the capacity for regeneration of a head is limited to the 
anterior third or fourth of the body, pieces from levels posterior 
to this failing to regenerate. Lillie noted that the pieces of Den- 
droccelum which were incapable of regenerating a head showed 
a marked difference in reactive power from those in which such 
regeneration was possible. On the other hand it is known (Loeb, 
'94,'99, Parker and Burnett, '00) that pieces of Planaria deprived 
of the cephalic ganglia react to stimuli in much the same manner 
as normal animals. In view of these differences in reactive power 
Lillie suggests that the stimulation of the normal movements may 
determine the fate of the undifferentiated mass of new tissue, the 
head failing to regenerate in the absence of the characteristic 
stimuli. This suggestion is, I think, an important one. 

As Lillie points out, Dendroccelum resembles in this respect 
the earthworm Allolohophora joetida, in which, according to 
Morgan ('97), regeneration of a head does not usually occur 
posterior to the fifteenth segment. In later work upon this form 
Morgan ('02) has discovered that the regeneration of the head 
appears to be closely connected with the presence of an anterior 
cut surface of the nerve cord, so that if two such surfaces are pre- 
sented by removing the head and then cutting out a small piece 
of the nerve cord a short distance posterior to the cut end, a head 
will regenerate from each of the cut surfaces. 

In my previous paper on Leptoplana (Child, '04) I suggested 
that the nervous stimuli in the region of a cut surface may exercise 
either directly or indirectly some influence upon the growth of new 
tissue from this region, and, moreover, that after removal of a 
part they may even be increased in intensity because of the more 
or less ineffectual attempts of the animal to perform the character- 
istic movements. 

The facts and conclusions cited, together with many others, 
such as the cases described by Herbst ('96a, '96b, '99) of the sub- 
stitution of an antenna-like organ for an eye in the absence of the 
optic ganglion from the eye-stalks of certain decapod Crustacea 

Studies on Regulation. 465 

and the extensive literature of the interesting, although at present 
somewhat confused question of the relation between the nervous 
system and the formation and development of the voluntary 
muscles (Herbst, '01, Neumann, '01, '03, Goldstein,^ '04) all 
afford evidence that there is a relation of some sort between the 
nervous system and the formation of certain structures, at least 
in some stages of development. 

Regarding the nature of this relation various opinions exist. 
The question as to the " trophic " influence of the nervous system is 
exceedingly obscure; the formative stimuli of Herbst and others 
are apparently regarded as entirely distinct from nervous func- 
tional stimuli. But that some relation exists between functional 
stimuli and the development and continued existence of certain 
structures cannot be doubted. 

The occurrence of regeneration in plants. Protozoa and other 
forms and in stages in which there is no visible diff^erentiation of 
the nervous system is of course no argument against the influence 
of the nervous system where it is present. For the development 
of the nervous system does not add anything to the protoplasm 
which is fundamentally diff^erent from what already exists there. 
The nervous system is simply a more or less highly diff^erentiated 
structure which accomplishes the transference and transforma- 
tion of stimuli, but in its absence some method of transference, 
however diff^use, must exist m the protoplasm. 

The following study of regeneration and other regulative phe- 
nomena in relation to the nervous system endeavors to present 
certain phases of the problem which seem to me important for the 
form considered. 

The figures are diagrammatic but are drawn from careful meas- 
urements in nearly all cases. In a number of cases the extent of 
the intestinal branches is indicated in the figure in a simple 
manner, no attempt being made to show the actual course of 
branches in particular individuals. The ganglia are drawn, 
where present, but the nerve cords are not indicated. The size 
of the pharynx is shown as exactly as possible: in most cases the 

^ Further references to the literature of this subject may be found in Goldstein's 

466 C. M. Child. 

genital ducts are indicated only by the "genital area" posterior 
to the pharynx as this was all that could be distinguished with 
certainty except when the ducts were filled with sexual products. 




The existence of a relation between the nervous system and both 
morphogenesis and the maintenance of form has been established 
or regarded as probable in various cases, some of which have 
already been mentioned. Some authors postulate the existence 
of special nervous "formative stimuli" and "trophic" nervous 
stimuli have been much discussed. But the relation between the 
nervous system and morphogenesis is of a problematic character, 
though the existence of a relation of some sort can scarcely be 
denied in many cases. 

This relation may conceivably be either direct or indirect. In 
the first case particular nervous stimuli of some sort are to be 
regarded as constituting in themselves formative factors. In the 
second case in consequence of certain nervous stimuli a particular 
part may be subjected to certain conditions which may be the 
formative factors, though themselves wholly different in character 
from nervous stimuli. The conditions connected with and 
resulting from a particular functional activity of a motor organ 
constitute a good example of the indirect relation. In general the 
functional activity of a motor organ is determined and controlled 
more or less completely by the nervous stimuli which affect it and 
adjoining regions. In consequence of these stimuli it functions 
more or less perfectly in a particular manner. The functional 
activity subjects the tissues of the part to a great variety of con- 
ditions, physical and chemical, external and internal, which, 
however, considered as a whole constitute a characteristic com- 
plex. Change in the kind or degree of functional activity is of 
course accompanied by changes in the complex of functional con- 
ditions to which the part is subjected. If these conditions play any 
part in the morphogenesis or form-maintenance a relation between 
the nervous system and form will appear to exist in such a case, but 

Studies on Regulation. 467 

upon analysis will be found to be indirect rather than direct. It 
is to be remembered, however, that even in cases of this kind a 
direct relation may also exist, /. e., the functional nervous stimuli 
themselves may conceivably exercise a direct influence of some 
sort upon the form. 

The question as to whether the complex of functional conditions 
exclusive of nervous stimuli may effect form is undoubtedly to be 
answered in the affirmative. The existence of a relation between 
these conditions and form has been established with more or less 
certainty for various cases by Roux and others. Little attempt 
has been made, however, to analyze these conditions or analysis 
has usually not proved very successful. The best examples of 
so-called functional structures are to be found in the tissues con- 
nected with movement. In these structures the arrangement of 
parts is very closely dependerit upon the conditions resulting from 
use of the organs in a characteristic manner. 

In most of the Turbellaria as well as in many other forms the 
whole body is more or less involved in the characteristic move- 
ments and thus becomes in a sense a complex motor organ. It is 
not improbable therefore that the various conditions to which the 
tissues are subjected in consequence of the characteristic move- 
ments are in certain cases important formative factors. I have 
already shown that such conditions are concerned in form 
regulation in Stenostoma and Leptoplana (Child, '02, '03a, '04). 
As will appear, the description in the following section of the rela- 
tion between the nervous system and motor activity in Lepto- 
plana is a necessary preliminary to the experiments to be presented. 

To what extent the functional conditions may constitute forma- 
tive factors in cases where motor activity is not concerned is a 
problem regarding which the data are at present few. I am in- 
clined to believe, however, that we shall find form to be essentially 
functional in very many cases where it is not at present so re- 
garded. Indeed in one sense all organic form is functional. 

Among the conditions resulting from functional activity me- 
chanical conditions are important. Their importance has been 
recognized in connection with the structure of bone, muscle and 
connective tissue, but I think they are important factors in many 

468 C. M. Child. 

other cases also. The formative effect of these conditions may 
conceivably be twofold; they may act as stimuli to growth or other 
changes, i. e., they may exert a "trophic" effect as Triepel and 
others have pointed out, or they may act in a direct mechanical 
manner, bringing about a particular arrangement of material. 
Both of these methods of action are important but the second 
has been much neglected in the analysis of formative conditions. 

The direct mechanical effect of pressure and tension upon the 
form of parts is, I believe, of great importance and may afford in 
some cases a simple explanation of phenomena which appear 
inexplicable from other points of view. A good case in point is 
the change of form called by Morgan "morphallaxis" in regu- 
lating pieces of Planaria and other Turbellaria. In the case of 
Stenostonia I have shown this change to be primarily mechanical 
in nature (Child, '02, '03a) and there is no doubt that in other 
forms the same factors are effective. In the case of Leptoplana 
the effect of mechanical conditions has already been shown in the 
preceding paper (Child, '04), and will be further considered in the 
present paper. 

But in many cases an indirect relation exists between the nerv- 
ous system and the mechanical conditions, as in the cases of Sten- 
ostoma and Leptoplana above mentioned, since the mechanical 
conditions effective here depend upon the use of the parts in a 
characteristic manner during locomotion. It is thus easy to see 
how factors, simple in themselves and entirely independent of the 
nervous system, may apparently stand in relation to it. The same 
is of course true with regard to other functional conditions as well 
as the mechanical factors. 

Even in cases where a direct relation between the nervous system 
and form may be shown to exist I see no necessity for assuming 
the existence of special "formative stimuli" or "trophic stimuli" 
as distinct from the functional stimuli. Moreover, extreme 
caution is necessary before concluding that a direct relation exists. 

The problem of organic form is undoubtedly the most complex 
and difficult of all biological problems. I do not think that the 
suggestions made here tend toward its simplification. The factors 
of organic form include all the activities of organic substance as 

Studies on Regulation. 469 

well as the environmental factors in varying degree. Indeed, in 
most cases, if not in all, we may regard organic form as the visible 
effect upon the protoplasm of functional activity in the widest sense, 
occurring in a given environment. But the basis of this functional 
activity is to be found in the composition of the protoplasm to- 
gether with environmental factors. I believe this distmction be- 
tween protoplasmic composition and organic form is important. 
In general the composition of the protoplasm determines — not 
form but functional activity of some sort, and m consequence of 
the internal or external conditions connected with the activity and 
produced by it form appears. We may say that morphological 
form is the visible expression of protoplasmic activity in a given 

If my experiments succeed in establishing for certain cases cer- 
tain definite factors in the complex of conditions upon which form 
depends, something has been gained, especially when we con- 
sider the vagueness or the anthropomorphic character of many 
hypotheses concerning form, and when we remember for instance 
that Driesch has made certain aspects of the problem of form the 
basis of his theory of the autonomy of the vital processes, while 
certain other authors hold that the problem is at present insoluble. 
If it has proven insoluble thus far I believe it is because of the 
methods employed rather than the nature of the problem. 


I. The Central Nervous System in Relation to Behavior. 

The characteristic movements of the normal animal (Child, 
'04) are coordinated in such manner that definite characteristic 
results are obtained: locomotion in a definite direction is possible 
and the motor reactions to various stimuli possess a definite 
character. Removal of the cephalic ganglia brings about a 
marked chang-e in the character of the movements. Pieces with- 
out the cephalic ganglia appear at first glance to be in great degree 
incapable of movement. Careful observation of the pieces shows, 
however, that they are capable of at least many of the character- 
istic movements of the species but that those movements are much 
less powerful and lack coordination. 

470 C. M. Child. 

But another important feature of the movements in the absence 
of the cephahc gangha must be noted, viz: that different pieces 
differ from each other in the degree of coordination, power, and 
frequency of their movements. Pieces from vs^hich the anterior 
end has been removed by a cut only a short distance posterior to 
the cephaHc gangha are capable of a somev^^hat greater degree of 
activity than those from which the anterior half or two-thirds of 
the body has been removed. In general it appears that the greater 
the remaining portion of the central nervous system the more 
complete the activity. 

We may consider first the case of a specimen from which the 
anterior end has been removed by a transverse cut two or three 
millimeters posterior to the cephalic ganglia. Such a piece is 
capable of locomotion but the advance is very slow and uniform. 
In my account of the normal movements (Child, '04) I called 
attention to the fact that locomotion in Leptoplana is accom- 
plished both by means of cilia and by muscular contraction, parts 
of the margin being extended and attached to the substratum and 
then undergoing contraction, thus dragging the body forward. 
The muscular factor is especially conspicuous after strong stimu- 
lation. In the specimen deprived of the cephalic ganglia, how- 
ever, progression is accomplished largely by means of cilia, hence 
the slow, uniform, gliding character of the movement. The 
specimen is apparently capable of performing all the muscular 
movements necessary for muscular locomotion but they appear to 
lack perfect coordination. Occasionally the piece seems to suc- 
ceed in using its muscles in some degree effectually, but it is 
probable that these instances are simply due to chance coincidence 
of particular muscular contractions. As the piece is more and 
more strongly stimulated the muscular contractions become more 
and more violent, although not coordinated, until finally the whole 
piece is involved in convulsive movements during which it may 
roll up and unroll or twist and squirm about, often turning over 
with ventral surface uppermost. 

Use of the posterior margins and posterior end of the body as 
organs of attachment occurs to some extent in these pieces. As 
the piece glides over the substratum parts of these regions can be 

Studies on Regulation. 471 

seen to attach and free themselves in the characteristic manner, 
though here the muscular pi,ay of the margins is much less marked. 
The piece as a whole adheres much less closely to the substratum, 
however, than the normal animal. It is not at all difficult to 
detach these pieces by means of a current of water from a pipette, 
while the normal animal adheres so closely that detachment by 
this method is often almost impossible. 

According to these observations pieces without the cephalic 
ganglia show both a quantitative and qualitative difference from 
normal animals as regards motor activities. All motor activities 
appear to be less intense than under ordinary conditions and the 
imperfect coordination in muscular movements alters the charac- 
ter of the movements very greatly. 

In these pieces the margins of the head, apparently the chief 
tactile organs, are of course absent and other parts of the body 
are less sensitive than these. Reaction to tactile stimulation of 
the lateral and other regions of the body is, however, less intense 
and definite than in pieces containing the cephalic ganglia. The 
eyes are also absent in these pieces and there is no marked re- 
action to light, though in a few cases, I thought I could observe 
some slight reaction (compare Parker and Burnett, '00). 

Individual differences in the behavior of pieces without ganglia 
are often observed even where the cuts removing the head were at 
the same level. Some pieces seem capable of more complete 
coordination than others, as is clearly seen, for example, by the 
rapidity with which they right themselves. These individual 
differences are of most frequent occurrence when the cut is not 
far from the ganglia and may be due to slight differences in level 
of the cut, one piece retaining some parts of the nervous system 
absent in others. Occasionally, however, they occur when the 
cut was some distance posterior to the ganglia, and in such cases 
must probably be ascribed to some structural or physiological 
difference of which at present we know little. The fact of the 
existence of such differences is however of interest as probably 
indicating the existence of marked variations of some kind in the 
nervous system. 

There seems to be some degree of correlation between size and 

472 CM. Child. 

the ability to perform coordinated movements in these pieces 
deprived of gangha. Of two pieces with anterior ends at the same 
level the longer seems to show a slightly greater degree of coordina- 
tion; it is sometimes able to advance more rapidly than the shorter 
piece and in general appears to be less completely helpless. The 
piece which has lost the greater posterior part of its body does not 
make up for this loss by greater use of the regenerating part to any 
such extent as does the piece with ganglia, but is simply more help- 
less than the piece which has lost only a small part. Exceptions are 
frequent but I think that a real difference does exist. It is neces- 
sary to distinguish two factors here, viz: the power of motor 
activity in general, i. e., the power of performing movements of 
any kind, and the power of coordinate functional activity. The 
smaller pieces usually appear to be more active than the larger 
but their activity seems to be less perfectly coordinated and so 
less effective as regards locomotion, etc. I am inclined to believe 
that the greater activity of the smaller pieces is connected with the 
loss of a large part of the body as is the case in similar pieces with 
ganglia, while on the other hand the lack of ability to coordinate 
is probably due to the small portion of the central nervous system 

The question as to whether the pieces deprived of the cephalic 
ganglia retain the power of "spontaneous" movement is some- 
what difficult to answer, since no sharp distinction can be made 
between spontaneous movements so-called and complex series of 
movements following particular stimuli; indeed in my opinion no 
distinction save one of degree exists. The pieces without cephalic 
ganglia are certainly much less active than normal animals, react 
more slowly and less strongly to stimuli and, as has been men- 
tioned, are unable to a large extent to coordinate their muscular 
movements. But even when apparently undisturbed such pieces 
are often found moving slowly about and performing indefinite 
muscular movements similar in character to those of normal 
animals but not correlated. I am inclined to believe that the loss 
of the cephalic ganglia means essentially the loss of the connections 
with the principal sense organs, i. e.^ the organs for the reception 
of stimuli, and the loss of a part of the more or less complex con- 

Studies on Regulation. 473 

ducting paths. This being the case we should expect to find less 
power of reaction to stimuli, less complexity and a lower degree of 
correlation in the movements. These are exactly the conditions 
that we do find. 

It is, I think, desirable to avoid the use of the word "sponta- 
neous" in this connection since the difference between sponta- 
neous and non-spontaneous movements seems to be merely one 
of degree of complexity and correlation or coordination. Re- 
moval of the principal paths by which stimuli enter and a part of 
the structures which connect these paths with other parts oi the 
nervous system must reduce the complexity of structure and con- 
sequently of the visible activities dependent upon this structure. 

Among these observations the most important point for the 
present consideration is the presence of the power of progressive 
locomotion in some degree in pieces deprived of the ganglia. 

Loeb ('94, '99) found that in the case of Thysanozoon loss of 
the power of progressive locomotion resulted from the removal of 
the cephalic ganglion. This is certainly not the case in Lepto- 
playia, and indeed experiments of my own upon Thysanozoon led 
me to the conclusion that even here the pieces without the ganglia 
still possessed some slight power of locomotion, though much less 
than that of the normal animal. In both Thysanozoon and Lepto- 
plana these pieces are capable of righting themselves after being 
turned over, but the change in position is much less rapid than in 
normal animals and frequently is accomplished only after re- 
peated attempts, or in some cases does not succeed at all, and the 
piece gradually becomes quiet. 

One other point of considerable interest must be considered. 
In Leptoplana I observed a marked difference in the power of 
locomotion and of coordination in general in pieces cut at different 
levels, the activity decreasing as the portion of the body removed 
with the cephalic ganglia increased. If, for instance, an individual 
was cut transversely two or three millimeters posterior to the 
ganglia the posterior piece was much more active and was capable 
of more rapid locomotion and more perfect coordination than a 
posterior piece obtained by a cut posterior to the middle of the 
body. In general the greater the distance between the cut and 

474 C. M. Child. 

the ganglia the less the activity and the more irregular and im- 
perfect are the movements. The difference between a piece ob- 
tained by a cut just posterior to the ganglia and one from the 
region posterior to the pharynx is striking. The latter scarcely 
reacts at all to stimuli, is almost v^holly incapable of progressive 
locomotion and rarely succeeds in righting itself — though this 
last may be due in part to the fact that such pieces are necessarily 
short— while the activity of the former is much greater in all 
respects though far below that of the normal animal. 

So far as I am aware the case of Dendroccelum mentioned by 
Lillie ('oi) is the only one in which observations of this kind have 
been made. In Detidroccelum Lillie found that posterior pieces 
obtained by section in the anterior third or fourth of the animal 
reacted to light like normal animals though more slowly, while 
pieces from levels posterior to this did not react. He does not 
mention any degree of reactive power corresponding to difference 
of level of the cut in pieces capable of reacting, but since other 
Turbellaria which I have observed resemble Leptoplana more or 
less closely, I am inclined to think that such a difference may 
possibly be present, though I have not had the opportunity of 
examining Dendrocoelum in sufficient numbers to decide this 

The differences in pieces from different levels are not due simply 
to differences in size, for a short piece from the region just pos- 
terior to the cephalic ganglia is much more active than a piece of 
the same size from the posterior region.- It is apparently not 
simply the amount of nerve tissue present that determines the 
degree of motor activity, but rather the quality of this tissue which 
differs in different regions. With our present knowledge of the 
nervous system only vague surmises as to the nature of this differ- 
ence are possible. It may be, at least in part, a difference in 
structural complexity, or the difference in the quantity of energy 
transformed by the stimuli or it may be something different from 
either of these: as a matter of fact the nerve cords in the Turbel- 
laria diminish in size toward the posterior end of the body. But 
the fact that a difference exists is important. My observations 
also indicate that the case is somewhat similar as regards the ceph- 

Studies on Regulation. 475 

alic ganglia: in general the motor activity falls further and further 
below the normal as the portion of the ganglia removed or injured 
increases. It is often difficult with the present technique of 
operation upon these forms to determine the extent of injury to 
the small ganglia, but notwithstanding this difficulty my observa- 
tions indicate very clearly that a relation exists between coordi- 
nated motor activity and the amount of ganglionic tissue present. 
When the cut passes through the middle of the ganglia both pieces 
separated behave essentially like normal animals, but when less 
than half of the ganglionic tissue remains intact the piece behaves 
much like specimens without ganglia and if the portions of the 
ganglia remaining are very small there is almost no motor activity 
except the ciliary movement, unless the piece is strongly stimu- 
lated. Since the ganglia are small and the difficulty of making a 
section in them at exactly the level desired is great, and since it is 
often difficult to determine after section just what parts of the 
ganglia remain, the results of these experiments are not exact. 
But the fact that the two pieces of an individual separated by a 
cut through the middle of the ganglia both behave like normal 
animals shows that the removal of half of the ganglionic tissue 
does not affect the behavior appreciably. Moreover, it makes 
no difference in such cases whether the cut is-longitudinal or other- 
wise. Anterior and posterior halves and right and left halves of 
the ganglia seem to be essentially alike in this respect. 

Pieces from the region anterior to the ganglia show almost no 
motor activity except that of the cilia, which continue to beat, and 
some degree of contraction after strong stimulation. Such pieces 
die in the course of two or three days. 

These relations between the various regions of the nerve cords 
and the cephalic ganglia and coordinated motor activity will be 
illustrated in the consideration of individual cases. The fact of 
the relation is of interest and indicates, in my opinion, that co- 
ordination is connected in these forms rather with a certain extent 
and structural complexity than with certain definite organs or 
centers. Certainly the cephalic ganglia are more important for 
motor activity and coordination than the other portions of the 
nervous system, but it is possible that their connection with the 

476 C. M. Child. 

chief sense organs, z. e., the paths by which more or less definitely 
localized stimuli enter the nervous system, is the primary factor 
in their predominance. 

With regard to the existence of "centers" in the nervous system 
I agree essentially M^ith Loeb ('99) and I think the relations above 
described support this view. Coordinated movements are the 
result of series of interrelations and exist after mutilation in the 
degree in which the interrelations remain intact or are reestab- 

The case of Leptoplaria as cited affords strong support to the 
view that the difference between "spontaneous" and "non- 
spontaneous" motor-activity is simply one of degree. Moreover, 
it is impossible to say that one part of the central nervous system 
in Leptoplana is necessarily connected with coordinated move- 
ment while another is not. It is rather the amount of nervous 
tissue — in all probability the completeness of the system of con- 
nections of parts — than the presence of any one portion which 
determines the results. 

2. The Relation Between the Central Nervous System and 
Posterior Regeneration. 

From Schultz's ('02) account of regeneration in Leptoplana 
atomata, it is evident that there is but little difference as regards 
regeneration between this species and L. tremellaris, but in the only 
case in which Schultz and I are really concerned with the same 
problem our interpretations of the facts differ widely. 

As regards the limits of regeneration in Leptoplana a brief 
preliminary statement will suffice here. Posterior regeneration 
from a cut surface is qualitatively complete at all levels posterior 
to the cephalic ganglia whether these are present or not and an- 
terior regeneration is complete only when the ganglia are present 
at least in large part, /. e., only anterior to them. In other words, 
regeneration of a head is impossible in the absence of the cephalic 
ganglia but posterior regeneration occurs whether they are present 
or not. In the absence of food the size of the new part is never as 
great as that of the part removed, but this is not of great impor- 

Studies on Regulation. 


The course of regeneration in the posterior direction from a 
level between the cephalic ganglia and the pharynx is illustrated 
in Figs. 1-5. On these figures the organs are indicated in a 
somewhat diagrammatic manner. The intestine is not drawn in 
the old parts, but the general distribution of its branches is indi- 
cated in the regenerated parts. Fig, i indicates the level of 
the cut and the shape ol the anterior end before section. After 
section the cut surface contracts and becomes concave posteriorly, 
and within two or three days new unpigmented tissue appears. In 
Fig. 2 the condition of the piece ten days after section is indi- 
cated. An outgrowth of new tissue tapering posteriorly is present, 
into which intestinal branches extend from the old part — and it 
may be mentioned in passing that the intestine in regenerated 

areas apparently always arises in connection with the old parts 
present. In the median line is a small ill-defined area which 
represents the developing pharynx. Fig. 3, sixteen days after 
section, shows a more advanced condition. The regenerated 
area is longer and the pharynx is distinct. From this time on a 
marked decrease in size occurs but the old part is much more 
affected than the new, as is indicated by Fig. 4 twenty-seven 
days after section. Here the new and old parts are of equal 
length, the new being longer, though perhaps not containing more 
material than in Fig. 3, and the old shorter. The pharynx 
has increased in size and beyond it a small clear area, which may 
be called the genital area, indicates that regeneration of the genital 
ducts is taking place. Fig. 5 shows a stage fifty-one days after 

478 CM. Child. 

section. The old parts have continued to decrease in size more 
rapidly than the new, but otherwise there is little difference be- 
tween this and the preceding stage. This piece remained alive 
during another month, the change in relative size of old and new 
parts continuing, together with reduction in size of the whole. 

The history of this piece is typical for posterior regeneration 
from this level of the body. Differences in the amount of regen- 
eration occur in different individuals, but in all cases regeneration 
may be said to be qualitatively complete in that the characteristic 
organs of the part removed are regenerated. As the size of the 
part removed decreases so in general does the amount of regenera- 
tion. The significance of this fact will be discussed more fully 
later. When parts of the pharynx or genital ducts are present 
the regeneration apparently always begins from the old part, but 
when such organs are wholly removed they are formed anew. In 
general the level at which regeneration occurs, the presence or 
absence of food, and individual differences affect the regeneration 
quantitatively but not qualitatively. 

a. Experiments on the Relation between the Cephalic Ganglia 
and Posterior Regeneration at Various Levels behind the 
Mention was made above of thefact that Leptoplana is not cap- 
able in any case of regenerating a head in the absence of the cephalic 
ganglia. This apparent dependence of anterior regeneration upon 
the cephalic ganglia has been established for a number of forms 
but the question as to the relation between the cephalic ganglia 
and posterior regeneration in the Turbellaria has received little 
attention. In order to examine this problem I prepared series of 
pieces as follows: a certain number of specimens of as nearly as 
possible the same size were cut at a given level and from half of 
these the cerebral ganglia were removed by a transverse cut just 
posterior to them; the regeneration of the two sets was then com- 
pared at stated intervals with respect to rapidity, amount, and 
quality of regeneration and the form of the new part. In several 
cases also series prepared for other purposes proved of value in 
this connection and could be compared with other pieces cut at 

Studies on Regulation. 


the same level which were not originally intended as controls for 
them. These experiments were performed during the winter 
when the temperature of the water was much lower than in sum- 
mer and the total amount of regeneration in the various cases is 
less than in summer experiments. In all cases the pieces were 
kept until regeneration ceased, in order that comparison of the 
total regeneration might be made. 

Series 73. Six pieces, each representing the region of the 
body between the cephalic ganglia and the pharynx, were 
obtained by the two transverse cuts indicated in Fig. 6. 

Series 82. Five pieces were obtained by transverse cuts just 
anterior to the pharynx (the lower line in Fig. 6) but the head 

and cephalic ganglia were left intact. This series was not origi- 
nally intended as a control for Series 73 consequently the inter- 
vals between examinations are somewhat different though not 
enough to prevent comparison. 

Fig. 7 shows the condition of the pieces of Series 73 eighteen 
days after section and Fig. 8 the posterior ends of the pieces of 
Series 82 fourteen days after section. In Series J^ the contraction 
of the cut surface is greater, the new tissue contains fewer intes- 
tinal branches, and the amount of the new tissue is somewhat less 
than in Series 82. 

The different pieces of each series were so closely similar that 
these two will serve as examples. 

48 o C. M. Child. 

In Figs. 9 and lo the condition of the pieces of Series 73 
thirty-eight days after section is indicated. In one of the pieces 
the new tissue showed the tapering form of Fig. 10, the other 
pieces resembhng Fig. 9. The former piece was capable of 
more rapid locomotion than the others. In all the pharynx and 
genital area are visible and an axial intestine with short branches 
extends down the middle of the pieces. 

The condition of the pieces of Series 82 thirty-four days after 
section is indicated in Fig. 11. Different pieces differed 
slightly as regards the length of the new tissue, but other differ- 
ences were not observed. Pharynx and genital area were present 
and the new tissue was well-filled with intestinal branches. 

The pieces of Series 82 differ markedly, however, from those of 
Series 73 in that the amount of regeneration is much greater in the 
Series 82, where the cephalic ganglia are present. Moreover, 
comparison of the Figs. 9 and 10 with Fig. 11 shows that the 
pharynx is longer and the intestinal branches much more abun- 
dant in Series 82. 

After this time there was no further advance in regeneration. 
The pieces of both series had already begun to decrease in size 
and continued to do so, but the decrease was somewhat more rapid 
in Series 82 than in Series 73. 

In these two series the differences seem to be wholly quantita- 
tive. The pieces in which the cephalic ganglia are intact regen- 
erate more rapidly, at least in the later stages; the amount of new 
tissue formed is greater; the pharynx is larger; and the intestinal 
branches are more numerous. 

Series 78. Five specimens were cut transversely through the 
middle of the pharynx (Fig. 12) the anterior part with head and 
cephalic ganglia intact being used. 

Series 79. Five specimens were cut at the same level as in 
Series 78 but in these the head was removed by a second cut just 
posterior to the ganglia (Fig. 12). 

Figs. 13 and 14 indicate the condition of the posterior ends in 
the two series fourteen days after section. In Fig. 14 (Series 79, 
without cephalic ganglia) the contraction of the cut surface is 
greater and the new tissue contains fewer intestinal branches 

Studies on Regulation. 


than in Fig. 13 (Series 78, ganglia present). There is no marked 
difference in the amount of regeneration. 

Thirty-four days after section the pieces of Series 78 have at- 
tained the condition represented in Figs. 15 and 16, and Figs. 
17 and 18 represent the condition of Series 79. Here, as in the 
preceding case there is a marked difference between the two 
series. In the series containing the cephaHc gangha, the amount 
of regeneration is greater, the posterior portion of the new pharynx 
has regenerated in the new tissue to a much larger extent, and 
intestinal branches fill the new tissue much more completely. 
Later stages afford no additional features of